Search Results

Search found 252 results on 11 pages for 'dimensionality reduction'.

Page 9/11 | < Previous Page | 5 6 7 8 9 10 11  | Next Page >

  • Comparing Apples and Pairs

    - by Tony Davis
    A recent study, High Costs and Negative Value of Pair Programming, by Capers Jones, pulls no punches in its assessment of the costs-to- benefits ratio of pair programming, two programmers working together, at a single computer, rather than separately. He implies that pair programming is a method rushed into production on a wave of enthusiasm for Agile or Extreme Programming, without any real regard for its effectiveness. Despite admitting that his data represented a far from complete study of the economics of pair programming, his conclusions were stark: it was 2.5 times more expensive, resulted in a 15% drop in productivity, and offered no significant quality benefits. The author provides a more scientific analysis than Jon Evans’ Pair Programming Considered Harmful, but the theme is the same. In terms of upfront-coding costs, pair programming is surely more expensive. The claim of productivity loss is dubious and contested by other studies. The third claim, though, did surprise me. The author’s data suggests that if both the pair and the individual programmers employ static code analysis and testing, then there is no measurable difference in the resulting code quality, in terms of defects per function point. In other words, pair programming incurs a massive extra cost for no tangible return in investment. There were, inevitably, many criticisms of his data and his conclusions, a few of which are persuasive. Firstly, that the driver/observer model of pair programming, on which the study bases its findings, is far from the most effective. For example, many find Ping-Pong pairing, based on use of test-driven development, far more productive. Secondly, that it doesn’t distinguish between “expert” and “novice” pair programmers– that is, independently of other programming skills, how skilled was an individual at pair programming. Thirdly, that his measure of quality is too narrow. This point rings true, certainly at Red Gate, where developers don’t pair program all the time, but use the method in short bursts, while tackling a tricky problem and needing a fresh perspective on the best approach, or more in-depth knowledge in a particular domain. All of them argue that pair programming, and collective code ownership, offers significant rewards, if not in terms of immediate “bug reduction”, then in removing the likelihood of single points of failure, and improving the overall quality and longer-term adaptability/maintainability of the design. There is also a massive learning benefit for both participants. One developer told me how he once worked in the same team over consecutive summers, the first time with no pair programming and the second time pair-programming two-thirds of the time, and described the increased rate of learning the second time as “phenomenal”. There are a great many theories on how we should develop software (Scrum, XP, Lean, etc.), but woefully little scientific research in their effectiveness. For a group that spends so much time crunching other people’s data, I wonder if developers spend enough time crunching data about themselves. Capers Jones’ data may be incomplete, but should cause a pause for thought, especially for any large IT departments, supporting commerce and industry, who are considering pair programming. It certainly shouldn’t discourage teams from exploring new ways of developing software, as long as they also think about how to gather hard data to gauge their effectiveness.

    Read the article

  • Managing Operational Risk of Financial Services Processes – part 2/2

    - by Sanjeev Sharma
    Normal 0 false false false EN-US X-NONE X-NONE MicrosoftInternetExplorer4 /* Style Definitions */ table.MsoNormalTable {mso-style-name:"Table Normal"; mso-tstyle-rowband-size:0; mso-tstyle-colband-size:0; mso-style-noshow:yes; mso-style-priority:99; mso-style-qformat:yes; mso-style-parent:""; mso-padding-alt:0in 5.4pt 0in 5.4pt; mso-para-margin:0in; mso-para-margin-bottom:.0001pt; mso-pagination:widow-orphan; font-size:10.0pt; font-family:"Calibri","sans-serif"; mso-bidi-font-family:"Times New Roman";} In my earlier blog post, I had described the factors that lead to compliance complexity of financial services processes. In this post, I will outline the business implications of the increasing process compliance complexity and the specific role of BPM in addressing the operational risk reduction objectives of regulatory compliance. First, let’s look at the business implications of increasing complexity of process compliance for financial institutions: · Increased time and cost of compliance due to duplication of effort in conforming to regulatory requirements due to process changes driven by evolving regulatory mandates, shifting business priorities or internal/external audit requirements · Delays in audit reporting due to quality issues in reconciling non-standard process KPIs and integrity concerns arising from the need to rely on multiple data sources for a given process Next, let’s consider some approaches to managing the operational risk of business processes. Financial institutions considering reducing operational risk of their processes, generally speaking, have two choices: · Rip-and-replace existing applications with new off-the shelf applications. · Extend capabilities of existing applications by modeling their data and process interactions, with other applications or user-channels, outside of the application boundary using BPM. The benefit of the first approach is that compliance with new regulatory requirements would be embedded within the boundaries of these applications. However pre-built compliance of any packaged application or custom-built application should not be mistaken as a one-shot fix for future compliance needs. The reason is that business needs and regulatory requirements inevitably out grow end-to-end capabilities of even the most comprehensive packaged or custom-built business application. Thus, processes that originally resided within the application will eventually spill outside the application boundary. It is precisely at such hand-offs between applications or between overlaying processes where vulnerabilities arise to unknown and accidental faults that potentially result in errors and lead to partial or total failure. The gist of the above argument is that processes which reside outside application boundaries, in other words, span multiple applications constitute a latent operational risk that spans the end-to-end value chain. For instance, distortion of data flowing from an account-opening application to a credit-rating system if left un-checked renders compliance with “KYC” policies void even when the “KYC” checklist was enforced at the time of data capture by the account-opening application. Oracle Business Process Management is enabling financial institutions to lower operational risk of such process ”gaps” for Financial Services processes including “Customer On-boarding”, “Quote-to-Contract”, “Deposit/Loan Origination”, “Trade Exceptions”, “Interest Claim Tracking” etc.. If you are faced with a similar challenge and need any guidance on the same feel free to drop me a note.

    Read the article

  • Getting Started with Cloud Computing

    - by juanlarios
    You’ve likely heard about how Office 365 and Windows Intune are great applications to get you started with Cloud Computing. Many of you emailed me asking for more info on what Cloud Computing is, including the distinction between "Public Cloud" and "Private Cloud". I want to address these questions and help you get started. Let's begin with a brief set of definitions and some places to find more info; however, an excellent place where you can always learn more about Cloud Computing is the Microsoft Virtual Academy. Public Cloud computing means that the infrastructure to run and manage the applications users are taking advantage of is run by someone else and not you. In other words, you do not buy the hardware or software to run your email or other services being used in your organization – that is done by someone else. Users simply connect to these services from their computers and you pay a monthly subscription fee for each user that is taking advantage of the service. Examples of Public Cloud services include Office 365, Windows Intune, Microsoft Dynamics CRM Online, Hotmail, and others. Private Cloud computing generally means that the hardware and software to run services used by your organization is run on your premises, with the ability for business groups to self-provision the services they need based on rules established by the IT department. Generally, Private Cloud implementations today are found in larger organizations but they are also viable for small and medium-sized businesses since they generally allow an automation of services and reduction in IT workloads when properly implemented. Having the right management tools, like System Center 2012, to implement and operate Private Cloud is important in order to be successful. So – how do you get started? The first step is to determine what makes the most sense to your organization. The nice thing is that you do not need to pick Public or Private Cloud – you can use elements of both where it makes sense for your business – the choice is yours. When you are ready to try and purchase Public Cloud technologies, the Microsoft Volume Licensing web site is a good place to find links to each of the online services. In particular, if you are interested in a trial for each service, you can visit the following pages: Office 365, CRM Online, Windows Intune, and Windows Azure. For Private Cloud technologies, start with some of the courses on Microsoft Virtual Academy and then download and install the Microsoft Private Cloud technologies including Windows Server 2008 R2 Hyper-V and System Center 2012 in your own environment and take it for a spin. Also, keep up to date with the Canadian IT Pro blog to learn about events Microsoft is delivering such as the IT Virtualization Boot Camps and more to get you started with these technologies hands on. Finally, I want to ask for your help to allow the team at Microsoft to continue to provide you what you need. Twice a year through something we call "The Global Relationship Study" – they reach out and contact you to see how they're doing and what Microsoft could do better. If you get an email from "Microsoft Feedback" with the subject line "Help Microsoft Focus on Customers and Partners" between March 5th and April 13th, please take a little time to tell them what you think. Cloud Computing Resources: Microsoft Server and Cloud Computing site – information on Microsoft's overall cloud strategy and products. Microsoft Virtual Academy – for free online training to help improve your IT skillset. Office 365 Trial/Info page – get more information or try it out for yourself. Office 365 Videos – see how businesses like yours have used Office 365 to transition to the cloud. Windows Intune Trial/Info – get more information or try it out for yourself. Microsoft Dynamics CRM Online page – information on trying and licensing Microsoft Dynamics CRM Online. Additional Resources You May Find Useful: Springboard Series Your destination for technical resources, free tools and expert guidance to ease the deployment and management of your Windows-based client infrastructure. TechNet Evaluation Center Try some of our latest Microsoft products for free, Like System Center 2012 Pre-Release Products, and evaluate them before you buy. AlignIT Manager Tech Talk Series A monthly streamed video series with a range of topics for both infrastructure and development managers. Ask questions and participate real-time or watch the on-demand recording. Tech·Days Online Discover what's next in technology and innovation with Tech·Days session recordings, hands-on labs and Tech·Days TV.

    Read the article

  • ProgrammingError when aggregating over an annotated & grouped Django ORM query

    - by ento
    I'm trying to construct a query to get the "average, maximum, minimum number of items purchased by a single user". The data source is this simple sales record table: class SalesRecord(models.Model): id = models.IntegerField(primary_key=True) user_id = models.IntegerField() product_code = models.CharField() price = models.IntegerField() created_at = models.DateTimeField() A new record is inserted into this table for every item purchased by a user. Here's my attempt at building the query: q = SalesRecord.objects.all() q = q.values('user_id').annotate( # group by user and count the # of records count=Count('id'), # (= # of items) ).order_by() result = q.aggregate(Max('count'), Min('count'), Avg('count')) When I try to execute the code, a ProgrammingError is raised at the last line: (1064, "You have an error in your SQL syntax; check the manual that corresponds to your MySQL server version for the right syntax to use near 'FROM (SELECT sales_records.user_id AS user_id, COUNT(sales_records.`' at line 1") Django's error screen shows that the SQL is SELECT FROM (SELECT `sales_records`.`player_id` AS `player_id`, COUNT(`sales_records`.`id`) AS `count` FROM `sales_records` WHERE (`sales_records`.`created_at` >= %s AND `sales_records`.`created_at` <= %s ) GROUP BY `sales_records`.`player_id` ORDER BY NULL) subquery It's not selecting anything! Can someone please show me the right way to do this? Hacking Django I've found that clearing the cache of selected fields in django.db.models.sql.BaseQuery.get_aggregation() seems to solve the problem. Though I'm not really sure this is a fix or a workaround. @@ -327,10 +327,13 @@ # Remove any aggregates marked for reduction from the subquery # and move them to the outer AggregateQuery. + self._aggregate_select_cache = None + self.aggregate_select_mask = None for alias, aggregate in self.aggregate_select.items(): if aggregate.is_summary: query.aggregate_select[alias] = aggregate - del obj.aggregate_select[alias] + if alias in obj.aggregate_select: + del obj.aggregate_select[alias] ... yields result: {'count__max': 267, 'count__avg': 26.2563, 'count__min': 1}

    Read the article

  • Difference in performance between Stax and DOM parsing

    - by Fazal
    I have been using DOM for a long time and as such DOM parsing performance wise has been pretty good. Even when dealing with XML of about 4-7 MB the parsing has been fast. The issue we face with DOM is the memory footprint which become huge as soon as we start dealing with large XMLs. Lately I tried moving to Stax (Streaming parsers for XML) which are supposed top be second generation parsers (reading about Stax it said its the fastest parser now). When I tried stax parser for large XML for about 4MB memory footprint definitely reduced drastically but time take to parse entire XML and create java object out of it increased almost by 5 times over DOM. I used sjsxp.jar implementation of Stax. I can deuce to some extent logically that performance may not be extremely good due to streaming nature of the parser but a reduction of 5 time (e.g. DOM takes about 8 seconds to build object for this XML, whereas Stax parsing took about 40 seconds on average) is definitely not going to be acceptable. Am I missing some point here completely as I am not able to come to terms with these performance numbers

    Read the article

  • Are there any radix/patricia/critbit trees for Python?

    - by Andrew Dalke
    I have about 10,000 words used as a set of inverted indices to about 500,000 documents. Both are normalized so the index is a mapping of integers (word id) to a set of integers (ids of documents which contain the word). My prototype uses Python's set as the obvious data type. When I do a search for a document I find the list of N search words and their corresponding N sets. I want to return the set of documents in the intersection of those N sets. Python's "intersect" method is implemented as a pairwise reduction. I think I can do better with a parallel search of sorted sets, so long as the library offers a fast way to get the next entry after i. I've been looking for something like that for some time. Years ago I wrote PyJudy but I no longer maintain it and I know how much work it would take to get it to a stage where I'm comfortable with it again. I would rather use someone else's well-tested code, and I would like one which supports fast serialization/deserialization. I can't find any, or at least not any with Python bindings. There is avltree which does what I want, but since even the pair-wise set merge take longer than I want, I suspect I want to have all my operations done in C/C++. Do you know of any radix/patricia/critbit tree libraries written as C/C++ extensions for Python? Failing that, what is the most appropriate library which I should wrap? The Judy Array site hasn't been updated in 6 years, with 1.0.5 released in May 2007. (Although it does build cleanly so perhaps It Just Works.)

    Read the article

  • OpenMP - running things in parallel and some in sequence within them

    - by Sayan Ghosh
    Hi, I have a scenario like: for (i = 0; i < n; i++) { for (j = 0; j < m; j++) { for (k = 0; k < x; k++) { val = 2*i + j + 4*k if (val != 0) { for(t = 0; t < l; t++) { someFunction((i + t) + someFunction(j + t) + k*t) } } } } } Considering this is block A, Now I have two more similar blocks in my code. I want to put them in parallel, so I used OpenMP pragmas. However I am not able to parallelize it, because I am a tad confused that which variables would be shared and private in this case. If the function call in the inner loop was an operation like sum += x, then I could have added a reduction clause. In general, how would one approach parallelizing a code using OpenMP, when we there is a nested for loop, and then another inner for loop doing the main operation. I tried declaring a parallel region, and then simply putting pragma fors before the blocks, but definitely I am missing a point there! Thanks, Sayan

    Read the article

  • Hierarchy of meaning

    - by asldkncvas
    I am looking for a method to build a hierarchy of words. Background: I am a "amateur" natural language processing enthusiast and right now one of the problems that I am interested in is determining the hierarchy of word semantics from a group of words. For example, if I have the set which contains a "super" representation of others, i.e. [cat, dog, monkey, animal, bird, ... ] I am interested to use any technique which would allow me to extract the word 'animal' which has the most meaningful and accurate representation of the other words inside this set. Note: they are NOT the same in meaning. cat != dog != monkey != animal BUT cat is a subset of animal and dog is a subset of animal. I know by now a lot of you will be telling me to use wordnet. Well, I will try to but I am actually interested in doing a very domain specific area which WordNet doesn't apply because: 1) Most words are not found in Wordnet 2) All the words are in another language; translation is possible but is to limited effect. another example would be: [ noise reduction, focal length, flash, functionality, .. ] so functionality includes everything in this set. I have also tried crawling wikipedia pages and applying some techniques on td-idf etc but wikipedia pages doesn't really do much either. Can someone possibly enlighten me as to what direction my research should go towards? (I could use anything)

    Read the article

  • OpenCL - incremental summation during compute

    - by user1721997
    I'm absolutelly novice in OpenCL programming. For my app. (molecular simulaton) I wrote a kernel for calculate intermolecular potential of lennard-jones liquid. In this kernel I need to compute cumulative value of the potential of all particles with one: __kernel void Molsim(__global const float* inmatrix, __global float* fi, const int c, const float r1, const float r2, const float r3, const float rc, const float epsilon, const float sigma, const float h1, const float h23) { float fi0; float fi1; float d; unsigned int i = get_global_id(0); //number of particles (typically 2000) if(c!=i) { // potential before particle movement d=sqrt(pow((0.5*h1-fabs(0.5*h1-fabs(inmatrix[c*3]-inmatrix[i*3]))),2.0)+pow((0.5*h23-fabs(0.5*h23-fabs(inmatrix[c*3+1]-inmatrix[i*3+1]))),2.0)+pow((0.5*h23-fabs(0.5*h23-fabs(inmatrix[c*3+2]-inmatrix[i*3+2]))),2.0)); if(d<rc) { fi0=4.0*epsilon*(pow(sigma/d,12.0)-pow(sigma/d,6.0)); } else { fi0=0; } // potential after particle movement d=sqrt(pow((0.5*h1-fabs(0.5*h1-fabs(r1-inmatrix[i*3]))),2.0)+pow((0.5*h23-fabs(0.5*h23-fabs(r2-inmatrix[i*3+1]))),2.0)+pow((0.5*h23-fabs(0.5*h23-fabs(r3-inmatrix[i*3+2]))),2.0)); if(d<rc) { fi1=4.0*epsilon*(pow(sigma/d,12.0)-pow(sigma/d,6.0)); } else { fi1=0; } // cumulative difference of potentials fi[0]+=fi1-fi0; } } My problem is in the line: fi[0]+=fi1-fi0;. In the one-element vector fi[0] are wrong results. I read something about sum reduction, but I do not know how to do it during the calculation. Exist any simple solution of my problem?

    Read the article

  • What limits scaling in this simple OpenMP program?

    - by Douglas B. Staple
    I'm trying to understand limits to parallelization on a 48-core system (4xAMD Opteron 6348, 2.8 Ghz, 12 cores per CPU). I wrote this tiny OpenMP code to test the speedup in what I thought would be the best possible situation (the task is embarrassingly parallel): // Compile with: gcc scaling.c -std=c99 -fopenmp -O3 #include <stdio.h> #include <stdint.h> int main(){ const uint64_t umin=1; const uint64_t umax=10000000000LL; double sum=0.; #pragma omp parallel for reduction(+:sum) for(uint64_t u=umin; u<umax; u++) sum+=1./u/u; printf("%e\n", sum); } I was surprised to find that the scaling is highly nonlinear. It takes about 2.9s for the code to run with 48 threads, 3.1s with 36 threads, 3.7s with 24 threads, 4.9s with 12 threads, and 57s for the code to run with 1 thread. Unfortunately I have to say that there is one process running on the computer using 100% of one core, so that might be affecting it. It's not my process, so I can't end it to test the difference, but somehow I doubt that's making the difference between a 19~20x speedup and the ideal 48x speedup. To make sure it wasn't an OpenMP issue, I ran two copies of the program at the same time with 24 threads each (one with umin=1, umax=5000000000, and the other with umin=5000000000, umax=10000000000). In that case both copies of the program finish after 2.9s, so it's exactly the same as running 48 threads with a single instance of the program. What's preventing linear scaling with this simple program?

    Read the article

  • Are there any downsides in using C++ for network daemons?

    - by badcat
    Hey guys! I've been writing a number of network daemons in different languages over the past years, and now I'm about to start a new project which requires a new custom implementation of a properitary network protocol. The said protocol is pretty simple - some basic JSON formatted messages which are transmitted in some basic frame wrapping to have clients know that a message arrived completely and is ready to be parsed. The daemon will need to handle a number of connections (about 200 at the same time) and do some management of them and pass messages along, like in a chat room. In the past I've been using mostly C++ to write my daemons. Often with the Qt4 framework (the network parts, not the GUI parts!), because that's what I also used for the rest of the projects and it was simple to do and very portable. This usually worked just fine, and I didn't have much trouble. Being a Linux administrator for a good while now, I noticed that most of the network daemons in the wild are written in plain C (of course some are written in other languages, too, but I get the feeling that 80% of the daemons are written in plain C). Now I wonder why that is. Is this due to a pure historic UNIX background (like KISS) or for plain portability or reduction of bloat? What are the reasons to not use C++ or any "higher level" languages for things like daemons? Thanks in advance! Update 1: For me using C++ usually is more convenient because of the fact that I have objects which have getter and setter methods and such. Plain C's "context" objects can be a real pain at some point - especially when you are used to object oriented programming. Yes, I'm aware that C++ is a superset of C, and that C code is basically C++. But that's not the point. ;)

    Read the article

  • Matrix Multiplication with C++ AMP

    - by Daniel Moth
    As part of our API tour of C++ AMP, we looked recently at parallel_for_each. I ended that post by saying we would revisit parallel_for_each after introducing array and array_view. Now is the time, so this is part 2 of parallel_for_each, and also a post that brings together everything we've seen until now. The code for serial and accelerated Consider a naïve (or brute force) serial implementation of matrix multiplication  0: void MatrixMultiplySerial(std::vector<float>& vC, const std::vector<float>& vA, const std::vector<float>& vB, int M, int N, int W) 1: { 2: for (int row = 0; row < M; row++) 3: { 4: for (int col = 0; col < N; col++) 5: { 6: float sum = 0.0f; 7: for(int i = 0; i < W; i++) 8: sum += vA[row * W + i] * vB[i * N + col]; 9: vC[row * N + col] = sum; 10: } 11: } 12: } We notice that each loop iteration is independent from each other and so can be parallelized. If in addition we have really large amounts of data, then this is a good candidate to offload to an accelerator. First, I'll just show you an example of what that code may look like with C++ AMP, and then we'll analyze it. It is assumed that you included at the top of your file #include <amp.h> 13: void MatrixMultiplySimple(std::vector<float>& vC, const std::vector<float>& vA, const std::vector<float>& vB, int M, int N, int W) 14: { 15: concurrency::array_view<const float,2> a(M, W, vA); 16: concurrency::array_view<const float,2> b(W, N, vB); 17: concurrency::array_view<concurrency::writeonly<float>,2> c(M, N, vC); 18: concurrency::parallel_for_each(c.grid, 19: [=](concurrency::index<2> idx) restrict(direct3d) { 20: int row = idx[0]; int col = idx[1]; 21: float sum = 0.0f; 22: for(int i = 0; i < W; i++) 23: sum += a(row, i) * b(i, col); 24: c[idx] = sum; 25: }); 26: } First a visual comparison, just for fun: The beginning and end is the same, i.e. lines 0,1,12 are identical to lines 13,14,26. The double nested loop (lines 2,3,4,5 and 10,11) has been transformed into a parallel_for_each call (18,19,20 and 25). The core algorithm (lines 6,7,8,9) is essentially the same (lines 21,22,23,24). We have extra lines in the C++ AMP version (15,16,17). Now let's dig in deeper. Using array_view and extent When we decided to convert this function to run on an accelerator, we knew we couldn't use the std::vector objects in the restrict(direct3d) function. So we had a choice of copying the data to the the concurrency::array<T,N> object, or wrapping the vector container (and hence its data) with a concurrency::array_view<T,N> object from amp.h – here we used the latter (lines 15,16,17). Now we can access the same data through the array_view objects (a and b) instead of the vector objects (vA and vB), and the added benefit is that we can capture the array_view objects in the lambda (lines 19-25) that we pass to the parallel_for_each call (line 18) and the data will get copied on demand for us to the accelerator. Note that line 15 (and ditto for 16 and 17) could have been written as two lines instead of one: extent<2> e(M, W); array_view<const float, 2> a(e, vA); In other words, we could have explicitly created the extent object instead of letting the array_view create it for us under the covers through the constructor overload we chose. The benefit of the extent object in this instance is that we can express that the data is indeed two dimensional, i.e a matrix. When we were using a vector object we could not do that, and instead we had to track via additional unrelated variables the dimensions of the matrix (i.e. with the integers M and W) – aren't you loving C++ AMP already? Note that the const before the float when creating a and b, will result in the underling data only being copied to the accelerator and not be copied back – a nice optimization. A similar thing is happening on line 17 when creating array_view c, where we have indicated that we do not need to copy the data to the accelerator, only copy it back. The kernel dispatch On line 18 we make the call to the C++ AMP entry point (parallel_for_each) to invoke our parallel loop or, as some may say, dispatch our kernel. The first argument we need to pass describes how many threads we want for this computation. For this algorithm we decided that we want exactly the same number of threads as the number of elements in the output matrix, i.e. in array_view c which will eventually update the vector vC. So each thread will compute exactly one result. Since the elements in c are organized in a 2-dimensional manner we can organize our threads in a two-dimensional manner too. We don't have to think too much about how to create the first argument (a grid) since the array_view object helpfully exposes that as a property. Note that instead of c.grid we could have written grid<2>(c.extent) or grid<2>(extent<2>(M, N)) – the result is the same in that we have specified M*N threads to execute our lambda. The second argument is a restrict(direct3d) lambda that accepts an index object. Since we elected to use a two-dimensional extent as the first argument of parallel_for_each, the index will also be two-dimensional and as covered in the previous posts it represents the thread ID, which in our case maps perfectly to the index of each element in the resulting array_view. The kernel itself The lambda body (lines 20-24), or as some may say, the kernel, is the code that will actually execute on the accelerator. It will be called by M*N threads and we can use those threads to index into the two input array_views (a,b) and write results into the output array_view ( c ). The four lines (21-24) are essentially identical to the four lines of the serial algorithm (6-9). The only difference is how we index into a,b,c versus how we index into vA,vB,vC. The code we wrote with C++ AMP is much nicer in its indexing, because the dimensionality is a first class concept, so you don't have to do funny arithmetic calculating the index of where the next row starts, which you have to do when working with vectors directly (since they store all the data in a flat manner). I skipped over describing line 20. Note that we didn't really need to read the two components of the index into temporary local variables. This mostly reflects my personal choice, in some algorithms to break down the index into local variables with names that make sense for the algorithm, i.e. in this case row and col. In other cases it may i,j,k or x,y,z, or M,N or whatever. Also note that we could have written line 24 as: c(idx[0], idx[1])=sum  or  c(row, col)=sum instead of the simpler c[idx]=sum Targeting a specific accelerator Imagine that we had more than one hardware accelerator on a system and we wanted to pick a specific one to execute this parallel loop on. So there would be some code like this anywhere before line 18: vector<accelerator> accs = MyFunctionThatChoosesSuitableAccelerators(); accelerator acc = accs[0]; …and then we would modify line 18 so we would be calling another overload of parallel_for_each that accepts an accelerator_view as the first argument, so it would become: concurrency::parallel_for_each(acc.default_view, c.grid, ...and the rest of your code remains the same… how simple is that? Comments about this post by Daniel Moth welcome at the original blog.

    Read the article

  • How to change radius of rounded rectangle in Photoshop

    - by MattDiPasquale
    At this point, I'm thinking of just using CSS 3, esp. since I'm a programmer, but I'd like to do this with Photoshop because I think it's nicer since I'm working with images anyway, among other reasons... Before I move on, my first question is: Is there a place like SuperUser for designers (or for Photoshop-like or questions)? What I Want: I want the icons on http://www.mattdipasquale.com/ to look like those on http://about.me/mattdipasquale. About.me has an outdated Twitter icon and does not have icons for GitHub, StackOverflow, etc. So, although I like the look of their icons, I want to be able to create these icons myself instead of using their versions. What I Have: I have different iphone icons, like the Facebook iPhone icon, Twitter iPhone icon, etc., that I got from iTunes, using Firebug to find the URL of the background image. I opened them up in Photoshop and pressed option + command + i to reduce the image size to 32px x 32px with Bicubic Sharper (best for reduction). I now have a square icon layer. Closing the Gap: In addition to the icon layer, I want to have a clipping-mask layer that will apply the 5px rounded-corners, 1px stroke, and 1px bevel. (Note: I just want to apply effects to the edges of the icon because the gloss and other effects are already encoded in the iTunes image. Also, I'm just guessing about the pixel values, but I want it to look good, like the icons on about.me.) What settings should I use for the blend options to make the icons look good, like iphone icons or those used by about.me? Why a Clipping Mask? The reason I want to use a clipping mask is that I want ease of reproducibility. I want to be able to apply the same styling to other square icon layers by simply replacing the square icon layer and then saving for web. If there is a better way to achieve such ease of reproducibility, please suggest it. I've seen Photoshop iPhone icon templates, but I couldn't figure out how to use them with my own images. Thanks! Matt

    Read the article

  • SSAS Compare: an intern’s journey

    - by Red Gate Software BI Tools Team
    About a month ago, David mentioned an intern working in the BI Tools Team. That intern happens to be me! In five weeks’ time, I’ll start my second year of Computer Science at the University of Cambridge and be a full-time student again, but for the past eight weeks, I’ve been living a completely different life. As Jon mentioned before, the teams here at Red Gate are small and everyone (including the interns!) is responsible for the product as a whole. I’ve attended planning sessions, UX tests, daily meetings, and everything else a full-time member of the team would; I had as much say in where we would go next with the product as anyone; I was able to see that what I was doing was an important part of the product from the feedback we got in the UX tests. All these things almost made me forget that this is just an internship and not my full-time job. First steps at Red Gate Being based in Cambridge, Red Gate has many Cambridge university graduates working for them. They also hire some Cambridge undergraduates for internships each summer. With its popularity with university graduates and its great working environment, Red Gate has managed to build up a great reputation. When I thought of doing an internship here in Cambridge, Red Gate just seemed to be the obvious choice for my first real work experience. On my first day at Red Gate, David, the lead developer for SSAS Compare, helped me settle in and explained what I’d be doing. My task was to improve the user experience of displaying differences between MDX scripts by syntax highlighting, script formatting, and improving the difference identification in the first place. David suggested how I should approach the problem, but left all the details and design decisions to me. That was when I realised how much independence and responsibility I’d have. What I’ve done If you launch the latest version of SSAS Compare and drill down to an MDX script difference, you can see the changes that have been made. In earlier versions, you could only see the scripts in plain text on both sides — either in black or grey, depending on whether they were the same or not. However, you couldn’t see exactly where the scripts were different, which was especially annoying when the two scripts were large – as they often are. Furthermore, if parts of the two scripts were formatted differently, they seemed to be different but were actually the same, which caused even more confusion and made it difficult to see where the differences were. All these issues have been fixed now. The two scripts are automatically formatted by the tool so that if two things are syntactically equivalent, they look the same – including case differences in keywords! The actual difference is highlighted in grey, which makes them easy to spot. The difference identification has been improved as well, so two scripts aren’t identified as different if there’s just a difference in meaningless whitespace characters, or when you have “select” on one side and “SELECT” on the other. We also have syntax highlighting, which makes it easier to read the scripts. How I did it In order to do the formatting properly, we decided to parse the MDX scripts. After some investigation into parser builders, I decided to go with the GOLD Parser builder and the bsn-goldparser .NET engine. GOLD Parser builder provides a fairly nice GUI to write, build, and test grammar in. We also liked the idea of separating the grammar building from parsing a text. The bsn-goldparser is one of many .NET engines for GOLD, and although it doesn’t support the newest features of GOLD Parser, it has “the ability to map semantic action classes to terminals or reduction rules, so that a completely functional semantic AST can be created directly without intermediate token AST representation, and without the need for glue code.” That makes it much easier for us to change the implementation in our program when we change the grammar. As bsn-goldparser is open source, and I wanted some more features in it, I contributed two new features which have now been merged to the project. Unfortunately, there wasn’t an MDX grammar written for GOLD already, so I had to write it myself. I was referencing MSDN to get the formal grammar specification, but the specification was all over the place, so it wasn’t that easy to implement and find. We’re aware that we don’t yet fully support all valid MDX, so sometimes you’ll just see the MDX script difference displayed the old way. In that case, there is some grammar construct we don’t yet recognise. If you come across something SSAS Compare doesn’t recognise, we’d love to hear about it so we can add it to our grammar. When some MDX script gets parsed, a tree is produced. That tree can then be processed into a list of inlines which deal with the correct formatting and can be outputted to the screen. Doing all this has led me to many new technologies and projects I haven’t worked with before. This was my first experience with C# and Visual Studio, although I have done things in Java before. I have learnt how to unit test with NUnit, how to do dependency injection with Ninject, how to source-control code with SVN and Mercurial, how to build with TeamCity, how to use GOLD, and many other things. What’s coming next Sadly, my internship comes to an end this week, so there will be less development on MDX difference view for a while. But the team is going to work on marking the differences better and making it consistent with difference indication in the top part of comparison window, and will keep adding support for more MDX grammar so you can see the differences easily in every comparison you make. So long! And maybe I’ll see you next summer!

    Read the article

  • Getting started with Oracle Database In-Memory Part III - Querying The IM Column Store

    - by Maria Colgan
    In my previous blog posts, I described how to install, enable, and populate the In-Memory column store (IM column store). This weeks post focuses on how data is accessed within the IM column store. Let’s take a simple query “What is the most expensive air-mail order we have received to date?” SELECT Max(lo_ordtotalprice) most_expensive_order FROM lineorderWHERE  lo_shipmode = 5; The LINEORDER table has been populated into the IM column store and since we have no alternative access paths (indexes or views) the execution plan for this query is a full table scan of the LINEORDER table. You will notice that the execution plan has a new set of keywords “IN MEMORY" in the access method description in the Operation column. These keywords indicate that the LINEORDER table has been marked for INMEMORY and we may use the IM column store in this query. What do I mean by “may use”? There are a small number of cases were we won’t use the IM column store even though the object has been marked INMEMORY. This is similar to how the keyword STORAGE is used on Exadata environments. You can confirm that the IM column store was actually used by examining the session level statistics, but more on that later. For now let's focus on how the data is accessed in the IM column store and why it’s faster to access the data in the new column format, for analytical queries, rather than the buffer cache. There are four main reasons why accessing the data in the IM column store is more efficient. 1. Access only the column data needed The IM column store only has to scan two columns – lo_shipmode and lo_ordtotalprice – to execute this query while the traditional row store or buffer cache has to scan all of the columns in each row of the LINEORDER table until it reaches both the lo_shipmode and the lo_ordtotalprice column. 2. Scan and filter data in it's compressed format When data is populated into the IM column it is automatically compressed using a new set of compression algorithms that allow WHERE clause predicates to be applied against the compressed formats. This means the volume of data scanned in the IM column store for our query will be far less than the same query in the buffer cache where it will scan the data in its uncompressed form, which could be 20X larger. 3. Prune out any unnecessary data within each column The fastest read you can execute is the read you don’t do. In the IM column store a further reduction in the amount of data accessed is possible due to the In-Memory Storage Indexes(IM storage indexes) that are automatically created and maintained on each of the columns in the IM column store. IM storage indexes allow data pruning to occur based on the filter predicates supplied in a SQL statement. An IM storage index keeps track of minimum and maximum values for each column in each of the In-Memory Compression Unit (IMCU). In our query the WHERE clause predicate is on the lo_shipmode column. The IM storage index on the lo_shipdate column is examined to determine if our specified column value 5 exist in any IMCU by comparing the value 5 to the minimum and maximum values maintained in the Storage Index. If the value 5 is outside the minimum and maximum range for an IMCU, the scan of that IMCU is avoided. For the IMCUs where the value 5 does fall within the min, max range, an additional level of data pruning is possible via the metadata dictionary created when dictionary-based compression is used on IMCU. The dictionary contains a list of the unique column values within the IMCU. Since we have an equality predicate we can easily determine if 5 is one of the distinct column values or not. The combination of the IM storage index and dictionary based pruning, enables us to only scan the necessary IMCUs. 4. Use SIMD to apply filter predicates For the IMCU that need to be scanned Oracle takes advantage of SIMD vector processing (Single Instruction processing Multiple Data values). Instead of evaluating each entry in the column one at a time, SIMD vector processing allows a set of column values to be evaluated together in a single CPU instruction. The column format used in the IM column store has been specifically designed to maximize the number of column entries that can be loaded into the vector registers on the CPU and evaluated in a single CPU instruction. SIMD vector processing enables the Oracle Database In-Memory to scan billion of rows per second per core versus the millions of rows per second per core scan rate that can be achieved in the buffer cache. I mentioned earlier in this post that in order to confirm the IM column store was used; we need to examine the session level statistics. You can monitor the session level statistics by querying the performance views v$mystat and v$statname. All of the statistics related to the In-Memory Column Store begin with IM. You can see the full list of these statistics by typing: display_name format a30 SELECT display_name FROM v$statname WHERE  display_name LIKE 'IM%'; If we check the session statistics after we execute our query the results would be as follow; SELECT Max(lo_ordtotalprice) most_expensive_order FROM lineorderWHERE lo_shipmode = 5; SELECT display_name FROM v$statname WHERE  display_name IN ('IM scan CUs columns accessed',                        'IM scan segments minmax eligible',                        'IM scan CUs pruned'); As you can see, only 2 IMCUs were accessed during the scan as the majority of the IMCUs (44) in the LINEORDER table were pruned out thanks to the storage index on the lo_shipmode column. In next weeks post I will describe how you can control which queries use the IM column store and which don't. +Maria Colgan

    Read the article

  • Windows Azure Recipe: Enterprise LOBs

    - by Clint Edmonson
    Enterprises are more and more dependent on their specialized internal Line of Business (LOB) applications than ever before. Naturally, the more software they leverage on-premises, the more infrastructure they need manage. It’s frequently the case that our customers simply can’t scale up their hardware purchases and operational staff as fast as internal demand for software requires. The result is that getting new or enhanced applications in the hands of business users becomes slower and more expensive every day. Being able to quickly deliver applications in a rapidly changing business environment while maintaining high standards of corporate security is a challenge that can be met right now by moving enterprise LOBs out into the cloud and leveraging Azure’s Access Control services. In fact, we’re seeing many of our customers (both large and small) see huge benefits from moving their web based business applications such as corporate help desks, expense tracking, travel portals, timesheets, and more to Windows Azure. Drivers Cost Reduction Time to market Security Solution Here’s a sketch of how many Windows Azure Enterprise LOBs are being architected and deployed: Ingredients Web Role – this will host the core of the application. Each web role is a virtual machine hosting an application written in ASP.NET (or optionally php, or node.js). The number of web roles can be scaled up or down as needed to handle peak and non-peak traffic loads. Many Java based applications are also being deployed to Windows Azure with a little more effort. Database – every modern web application needs to store data. SQL Azure databases look and act exactly like their on-premise siblings but are fault tolerant and have data redundancy built in. Access Control – this service is necessary to establish federated identity between the cloud hosted application and an enterprise’s corporate network. It works in conjunction with a secure token service (STS) that is hosted on-premises to establish the corporate user’s identity and credentials. The source code for an on-premises STS is provided in the Windows Azure training kit and merely needs to be customized for the corporate environment and published on a publicly accessible corporate web site. Once set up, corporate users see a near seamless single sign-on experience. Reporting – businesses live and die by their reports and SQL Azure Reporting, based on SQL Server Reporting 2008 R2, can serve up reports with tables, charts, maps, gauges, and more. These reports can be accessed from the Windows Azure Portal, through a web browser, or directly from applications. Service Bus (optional) – if deep integration with other applications and systems is needed, the service bus is the answer. It enables secure service layer communication between applications hosted behind firewalls in on-premises or partner datacenters and applications hosted inside Windows Azure. The Service Bus provides the ability to securely expose just the information and services that are necessary to create a simpler, more secure architecture than opening up a full blown VPN. Data Sync (optional) – in cases where the data stored in the cloud needs to be shared internally, establishing a secure one-way or two-way data-sync connection between the on-premises and off-premises databases is a perfect option. It can be very granular, allowing us to specify exactly what tables and columns to synchronize, setup filters to sync only a subset of rows, set the conflict resolution policy for two-way sync, and specify how frequently data should be synchronized Training Labs These links point to online Windows Azure training labs where you can learn more about the individual ingredients described above. (Note: The entire Windows Azure Training Kit can also be downloaded for offline use.) Windows Azure (16 labs) Windows Azure is an internet-scale cloud computing and services platform hosted in Microsoft data centers, which provides an operating system and a set of developer services which can be used individually or together. It gives developers the choice to build web applications; applications running on connected devices, PCs, or servers; or hybrid solutions offering the best of both worlds. New or enhanced applications can be built using existing skills with the Visual Studio development environment and the .NET Framework. With its standards-based and interoperable approach, the services platform supports multiple internet protocols, including HTTP, REST, SOAP, and plain XML SQL Azure (7 labs) Microsoft SQL Azure delivers on the Microsoft Data Platform vision of extending the SQL Server capabilities to the cloud as web-based services, enabling you to store structured, semi-structured, and unstructured data. Windows Azure Services (9 labs) As applications collaborate across organizational boundaries, ensuring secure transactions across disparate security domains is crucial but difficult to implement. Windows Azure Services provides hosted authentication and access control using powerful, secure, standards-based infrastructure. See my Windows Azure Resource Guide for more guidance on how to get started, including links web portals, training kits, samples, and blogs related to Windows Azure.

    Read the article

  • The JavaOne 2012 Sunday Technical Keynote

    - by Janice J. Heiss
    At the JavaOne 2012 Sunday Technical Keynote, held at the Masonic Auditorium, Mark Reinhold, Chief Architect, Java Platform Group, stated that they were going to do things a bit differently--"rather than 20 minutes of SE, and 20 minutes of FX, and 20 minutes of EE, we're going to mix it up a little," he said. "For much of it, we're going to be showing a single application, to show off some of the great work that's been done in the last year, and how Java can scale well--from the cloud all the way down to some very small embedded devices, and how JavaFX scales right along with it."Richard Bair and Jasper Potts from the JavaFX team demonstrated a JavaOne schedule builder application with impressive navigation, animation, pop-overs, and transitions. They noted that the application runs seamlessly on either Windows or Macs, running Java 7. They then ran the same application on an Ubuntu Linux machine--"it just works," said Blair.The JavaFX duo next put the recently released JavaFX Scene Builder through its paces -- dragging and dropping various image assets to build the application's UI, then fine tuning a CSS file for the finished look and feel. Among many other new features, in the past six months, JavaFX has released support for H.264 and HTTP live streaming, "so you can get all the real media playing inside your JavaFX application," said Bair. And in their developer preview builds of JavaFX 8, they've now split the rendering thread from the UI thread, to better take advantage of multi-core architectures.Next, Brian Goetz, Java Language Architect, explored language and library features planned for Java SE 8, including Lambda expressions and better parallel libraries. These feature changes both simplify code and free-up libraries to more effectively use parallelism. "It's currently still a lot of work to convert an application from serial to parallel," noted Goetz.Reinhold had previously boasted of Java scaling down to "small embedded devices," so Blair and Potts next ran their schedule builder application on a small embedded PandaBoard system with an OMAP4 chip set. Connected to a touch screen, the embedded board ran the same JavaFX application previously seen on the desktop systems, but now running on Java SE Embedded. (The systems can be seen and tried at four of the nearby JavaOne hotels.) Bob Vandette, Java Embedded Architect, then displayed a $25 Rasberry Pi ARM-based system running Java SE Embedded, noting the even greater need for the platform independence of Java in such highly varied embedded processor spaces. Reinhold and Vandetta discussed Project Jigsaw, the planned modularization of the Java SE platform, and its deferral from the Java 8 release to Java 9. Reinhold demonstrated the promise of Jigsaw by running a modularized demo version of the earlier schedule builder application on the resource constrained Rasberry Pi system--although the demo gods were not smiling down, and the application ultimately crashed.Reinhold urged developers to become involved in the Java 8 development process--getting the weekly builds, trying out their current code, and trying out the new features:http://openjdk.java.net/projects/jdk8http://openjdk.java.net/projects/jdk8/spechttp://jdk8.java.netFrom there, Arun Gupta explored Java EE. The primary themes of Java EE 7, Gupta stated, will be greater productivity, and HTML 5 functionality (WebSocket, JSON, and HTML 5 forms). Part of the planned productivity increase of the release will come from a reduction in writing boilerplate code--through the widespread use of dependency injection in the platform, along with default data sources and default connection factories. Gupta noted the inclusion of JAX-RS in the web profile, the changes and improvements found in JMS 2.0, as well as enhancements to Java EE 7 in terms of JPA 2.1 and EJB 3.2. GlassFish 4 is the reference implementation of Java EE 7, and currently includes WebSocket, JSON, JAX-RS 2.0, JMS 2.0, and more. The final release is targeted for Q2, 2013. Looking forward to Java EE 8, Gupta explored how the platform will provide multi-tenancy for applications, modularity based on Jigsaw, and cloud architecture. Meanwhile, Project Avatar is the group's incubator project for designing an end-to-end framework for building HTML 5 applications. Santiago Pericas-Geertsen joined Gupta to demonstrate their "Angry Bids" auction/live-bid/chat application using many of the enhancements of Java EE 7, along with an Avatar HTML 5 infrastructure, and running on the GlassFish reference implementation.Finally, Gupta covered Project Easel, an advanced tooling capability in NetBeans for HTML5. John Ceccarelli, NetBeans Engineering Director, joined Gupta to demonstrate creating an HTML 5 project from within NetBeans--formatting the project for both desktop and smartphone implementations. Ceccarelli noted that NetBeans 7.3 beta will be released later this week, and will include support for creating such HTML 5 project types. Gupta directed conference attendees to: http://glassfish.org/javaone2012 for everything about Java EE and GlassFish at JavaOne 2012.

    Read the article

  • Educause Top-Ten IT Issues - the most change in a decade or more

    - by user739873
    The Education IT Issue Panel has released the 2012 top-ten issues facing higher education IT leadership, and instead of the customary reshuffling of the same deck, the issues reflect much of the tumult and dynamism facing higher education generally.  I find it interesting (and encouraging) that at the top of this year's list is "Updating IT Professionals' Skills and Roles to Accommodate Emerging Technologies and Changing IT Management and Service Delivery Models."  This reflects, in my view, the realization that higher education IT must change in order to fully realize the potential for transforming the institution, and therefore it's people must learn new skills, understand and accept new ways of solving problems, and not be tied down by past practices or institutional inertia. What follows in the remaining 9 top issues all speak, in some form or fashion, to the need for dramatic change, but not just in the areas of "funding IT" (code for cost containment or reduction), but rather the need to increase effectiveness and efficiency of the institution through the use of technology—leveraging the wave of BYOD (Bring Your Own Device) to the institution's advantage, rather than viewing it as a threat and a problem to be contained. Although it's #10 of 10, IT Governance (and establishment and implementation of the governance model throughout the institution) is key to effectively acting upon many of the preceding issues in this year's list.  In the majority of cases, technology exists to meet the needs and requirements to effectively address many of the challenges outlined in top-ten issues list. Which brings me to my next point. Although I try not to sound too much like an Oracle commercial in these (all too infrequent) blog posts, I can't help but point out how much confluence there is between several of the top issues this year and what my colleagues and I have been evangelizing for some time. Starting from the bottom of the list up: 1) I'm gratified that research and the IT challenges it presents has made the cut.  Big Data (or Large Data as it's phased in the report) is rapidly going to overwhelm much of what exists today even at our most prepared and well-equipped research universities.  Combine large data with the significantly more stringent requirements around data preservation, archiving, sharing, curation, etc. coming from granting agencies like NSF, and you have the brewing storm that could result in a lot of "one-off" solutions to a problem that could very well be addressed collectively and "at scale."   2) Transformative effects of IT – while I see more and more examples of this, there is still much more that can be achieved. My experience tells me that culture (as the report indicates or at least poses the question) gets in the way more than technology not being up to task.  We spend too much time on "context" and not "core," and get lost in the weeds on the journey to truly transforming the institution with technology. 3) Analytics as a key element in improving various institutional outcomes.  In our work around Student Success, we see predictive "academic" analytics as essential to getting in front of the Student Success issue, regardless of how an institution or collections of institutions defines success.  Analytics must be part of the fabric of the key academic enterprise applications, not a bolt-on.  We will spend a significant amount of time on this topic during our semi-annual Education Industry Strategy Council meeting in Washington, D.C. later this month. 4) Cloud strategy for the broad range of applications in the academic enterprise.  Some of the recent work by Casey Green at the Campus Computing Survey would seem to indicate that there is movement in this area but mostly in what has been termed "below the campus" application areas such as collaboration tools, recruiting, and alumni relations.  It's time to get serious about sourcing elements of mature applications like student information systems, HR, Finance, etc. leveraging a model other than traditional on-campus custom. I've only selected a few areas of the list to highlight, but the unifying theme here (and this is where I run the risk of sounding like an Oracle commercial) is that these lofty goals cry out for partners that can bring economies of scale to bear on the problems married with a deep understanding of the nuances unique to higher education.  In a recent piece in Educause Review on Student Information Systems, the author points out that "best of breed is back". Unfortunately I am compelled to point out that best of breed is a large part of the reason we have made as little progress as we have as an industry in advancing some of the causes outlined above.  Don't confuse "integrated" and "full stack" for vendor lock-in.  The best-of-breed market forces that Ron points to ensure that solutions have to be "integratable" or they don't survive in the marketplace. However, by leveraging the efficiencies afforded by adopting solutions that are pre-integrated (and possibly metered out as a service) allows us to shed unnecessary costs – as difficult as these decisions are to make and to drive throughout the organization. Cole

    Read the article

  • Gamification = -10#/3mo

    - by erikanollwebb
    One of the purposes of gamification of anything is to see if you can modify the behavior of the user. In the enterprise, that might mean getting sales people to enter more information into a CRM system, encouraging employees to update their HR records, motivating people to participate in forums and discussions, or process invoices more quickly.  Wikipedia defines behavior modification as "the traditional term for the use of empirically demonstrated behavior change techniques to increase or decrease the frequency of behaviors, such as altering an individual's behaviors and reactions to stimuli through positive and negative reinforcement of adaptive behavior and/or the reduction of behavior through its extinction, punishment and/or satiation."  Gamification is just a way to modify someone's behavior using game mechanics. And the magic question is always whether it works. So I thought I would present my own little experiment from the last few months.  This spring, I upgraded to a Samsung Galaxy 4.  It's a pretty sweet phone in many ways, but one of the little extras I discovered was a built in app called S Health. S Health is an app that you can use to track calories, weight, exercise and it has a built in pedometer. I looked at it when I got the phone, but assumed you had to turn it on to use it so I didn't look at it much.  But sometime in July, I realized that in fact, it just ran in the background and was quietly tracking my steps, with a goal of 10,000 per day.  10,000 steps per day is this magic number recommended by the Surgeon General and the American Heart Association.  Dr. Oz pushes it as the goal for daily exercise.  It's about 5 miles of walking. I'm generally not the kind of person who always has my phone with me.  I leave it in my purse and pull it out when I need it.  But then I realized that meant I wasn't getting a good measure of my steps.  I decided to do a little experiment, and carry it with me as much as possible for a week.  That's when I discovered the gamification that changed my life over the last 3 months.  When I hit 10,000 steps, the app jingled out a little "success!" tune and I got a badge.  I was hooked.  I started carrying my phone.  I started making sure I had shoes I could walk in with me.  I started walking at lunch time, because I realized how often I sat at my desk for 8-10 hours every day without moving.  I started pestering my husband to walk with me after work because I hadn't hit my 10,000 yet, leading him at one point to say "I'm not as much a slave to that badge as you are!"  I started looking at parking lots differently.  Can't get a space up close?  No worries, just that many steps toward my 10,000.  I even tried to see if there was a second power user level at 15,000 or 20,000 (*sadly, no).  If I was close at the end of the day, I have done laps around my house until I got my badge.  I have walked around the block one more time to get my badge.  I have mentally chastised myself when I forgot to put my phone in my pocket because I don't know how many steps I got.  The badge below I got when my boss and I were in New York City and we walked around the block of our hotel just to watch the badge pop up. There are a bunch of tools out on the market now that have similar ideas for helping you to track your exercise, make it social.  There are apps (my favorite is still Zombies, Run!).  You could buy a FitBit or UP by Jawbone.   Interactive fitness makes the Expresso stationary bike with built in video games.  All designed to help you be more aware of your activity and keep you engaged and motivated.  And the idea is to help you change your behavior. I know someone who would spend extra time and work hard on the Expresso because he had built up strategies for how to kill the most dragons while he was riding to get more points.  When the machine broke down, he didn't ride a different bike because it just wasn't that interesting. But for me, just the simple jingle and badge have been all I needed.  I admit, I still giggle gleefully when I hear the tune sing out from my pocket. After a few weeks, I noticed I had dropped a few pounds.  Not a lot, just 2-3.  But then I was really hooked.  I started making a point both to eat a little less and hit 10,000 steps as much as I could.  I bemoaned that during the floods in Boulder, I wasn't hitting my 10,000 steps.  And now, a few months later, I'm almost 10 lbs lighter. All for 1 badge a day. So yes, simple gamification can increase motivation and engagement.  And that can lead to changes in behavior.  Now the job is to apply that to the enterprise space in a meaningful and engaging way. 

    Read the article

  • SOA, Governance, and Drugs

    Why is IT governance important in service oriented architecture (SOA)? IT Governance provides a framework for making appropriate decisions based on company guidelines and accepted standards. This framework also outlines each stakeholder’s responsibilities and authority when making important architectural or design decisions. Furthermore, this framework of governance defines parameters and constraints that are used to give context and perspective when making decisions. The use of governance as it applies to SOA ensures that specific design principles and patterns are used when developing and maintaining services. When governance is consistently applied systems the following benefits are achieved according to Anne Thomas Manes in 2010. Governance makes sure that services conform to standard interface patterns, common data modeling practices, and promotes the incorporation of existing system functionality by building on top of other available services across a system. Governance defines development standards based on proven design principles and patterns that promote reuse and composition. Governance provides developers a set of proven design principles, standards and practices that promote the reduction in system based component dependencies.  By following these guidelines, individual components will be easier to maintain. For me personally, I am a fan of IT governance, and feel that it valuable part of any corporate IT department. However, depending on how it is implemented can really affect the value of using IT governance.  Companies need to find a way to ensure that governance does not become extreme in its policies and procedures. I know for me personally, I would really dislike working under a completely totalitarian or laissez-faire version of governance. Developers need to be able to be creative in their designs and too much governance can really impede the design process and prevent the most optimal design from being developed. On the other hand, with no governance enforced, no standards will be followed and accepted design patterns will be ignored. I have personally had to spend a lot of time working on this particular scenario and I have found that the concept of code reuse and composition is almost nonexistent.  Based on this, too much time and money is wasted on redeveloping existing aspects of an application that already exist within the system as a whole. I think moving forward we will see a staggered form of IT governance, regardless if it is for SOA or IT in general.  Depending on the size of a company and the size of its IT department,  I can see IT governance as a layered approach in that the top layer will be defined by enterprise architects that focus on abstract concepts pertaining to high level design, general  guidelines, acceptable best practices, and recommended design patterns.  The next layer will be defined by solution architects or department managers that further expand on abstracted guidelines defined by the enterprise architects. This layer will contain further definitions as to when various design patterns, coding standards, and best practices are to be applied based on the context of the solutions that are being developed by the department. The final layer will be defined by the system designer or a solutions architect assed to a project in that they will define what design patterns will be used in a solution, naming conventions, as well as outline how a system will function based on the best practices defined by the previous layers. This layered approach allows for IT departments to be flexible in that system designers have creative leeway in designing solutions to meet the needs of the business, but they must operate within the confines of the abstracted IT governance guidelines.  A real world example of this can be seen in the United States as it pertains to governance of the people in that the US government defines rules and regulations in the abstract and then the state governments take these guidelines and applies them based on the will of the people in each individual state. Furthermore, the county or city governments are the ones that actually enforce these rules based on how they are interpreted by local community.  To further define my example, the United States government defines that marijuana is illegal. Each individual state has the option to determine this regulation as it wishes in that the state of Florida determines that all uses of the drug are illegal, but the state of California legally allows the use of marijuana for medicinal purposes only. Based on these accepted practices each local government enforces these rules in that a police officer will arrest anyone in the state of Florida for having this drug on them if they walk down the street, but in California if a person has a medical prescription for the drug they will not get arrested.  REFERENCESThomas Manes, Anne. (2010). Understanding SOA Governance: http://www.soamag.com/I40/0610-2.php

    Read the article

  • Python bindings for C++ code using OpenCV giving segmentation fault

    - by lightalchemist
    I'm trying to write a python wrapper for some C++ code that make use of OpenCV but I'm having difficulties returning the result, which is a OpenCV C++ Mat object, to the python interpreter. I've looked at OpenCV's source and found the file cv2.cpp which has conversions functions to perform conversions to and fro between PyObject* and OpenCV's Mat. I made use of those conversions functions but got a segmentation fault when I tried to use them. I basically need some suggestions/sample code/online references on how to interface python and C++ code that make use of OpenCV, specifically with the ability to return OpenCV's C++ Mat to the python interpreter or perhaps suggestions on how/where to start investigating the cause of the segmentation fault. Currently I'm using Boost Python to wrap the code. Thanks in advance to any replies. The relevant code: // This is the function that is giving the segmentation fault. PyObject* ABC::doSomething(PyObject* image) { Mat m; pyopencv_to(image, m); // This line gives segmentation fault. // Some code to create cppObj from CPP library that uses OpenCV cv::Mat processedImage = cppObj->align(m); return pyopencv_from(processedImage); } The conversion functions taken from OpenCV's source follows. The conversion code gives segmentation fault at the commented line with "if (!PyArray_Check(o)) ...". static int pyopencv_to(const PyObject* o, Mat& m, const char* name = "<unknown>", bool allowND=true) { if(!o || o == Py_None) { if( !m.data ) m.allocator = &g_numpyAllocator; return true; } if( !PyArray_Check(o) ) // Segmentation fault inside PyArray_Check(o) { failmsg("%s is not a numpy array", name); return false; } int typenum = PyArray_TYPE(o); int type = typenum == NPY_UBYTE ? CV_8U : typenum == NPY_BYTE ? CV_8S : typenum == NPY_USHORT ? CV_16U : typenum == NPY_SHORT ? CV_16S : typenum == NPY_INT || typenum == NPY_LONG ? CV_32S : typenum == NPY_FLOAT ? CV_32F : typenum == NPY_DOUBLE ? CV_64F : -1; if( type < 0 ) { failmsg("%s data type = %d is not supported", name, typenum); return false; } int ndims = PyArray_NDIM(o); if(ndims >= CV_MAX_DIM) { failmsg("%s dimensionality (=%d) is too high", name, ndims); return false; } int size[CV_MAX_DIM+1]; size_t step[CV_MAX_DIM+1], elemsize = CV_ELEM_SIZE1(type); const npy_intp* _sizes = PyArray_DIMS(o); const npy_intp* _strides = PyArray_STRIDES(o); bool transposed = false; for(int i = 0; i < ndims; i++) { size[i] = (int)_sizes[i]; step[i] = (size_t)_strides[i]; } if( ndims == 0 || step[ndims-1] > elemsize ) { size[ndims] = 1; step[ndims] = elemsize; ndims++; } if( ndims >= 2 && step[0] < step[1] ) { std::swap(size[0], size[1]); std::swap(step[0], step[1]); transposed = true; } if( ndims == 3 && size[2] <= CV_CN_MAX && step[1] == elemsize*size[2] ) { ndims--; type |= CV_MAKETYPE(0, size[2]); } if( ndims > 2 && !allowND ) { failmsg("%s has more than 2 dimensions", name); return false; } m = Mat(ndims, size, type, PyArray_DATA(o), step); if( m.data ) { m.refcount = refcountFromPyObject(o); m.addref(); // protect the original numpy array from deallocation // (since Mat destructor will decrement the reference counter) }; m.allocator = &g_numpyAllocator; if( transposed ) { Mat tmp; tmp.allocator = &g_numpyAllocator; transpose(m, tmp); m = tmp; } return true; } static PyObject* pyopencv_from(const Mat& m) { if( !m.data ) Py_RETURN_NONE; Mat temp, *p = (Mat*)&m; if(!p->refcount || p->allocator != &g_numpyAllocator) { temp.allocator = &g_numpyAllocator; m.copyTo(temp); p = &temp; } p->addref(); return pyObjectFromRefcount(p->refcount); } My python test program: import pysomemodule # My python wrapped library. import cv2 def main(): myobj = pysomemodule.ABC("faces.train") # Create python object. This works. image = cv2.imread('61.jpg') processedImage = myobj.doSomething(image) cv2.imshow("test", processedImage) cv2.waitKey() if __name__ == "__main__": main()

    Read the article

  • Is the recent trend toward widescreen (16:9) computer monitors a plus or minus for programmers?

    - by DanM
    It's almost gotten to the point where you can't buy a conventional (4:3) monitor anymore. Pretty much everything is widescreen. This is fine for watching movies or TV, but is it good or bad for programming? My initial thoughts on the issue are that widescreens are a net negative for programmers. Here are some of the disadvantages I see: Poor space utiliziation One disadvantage of widescreens you can't argue with is that they offer poor space utilization for the amount of total pixels you get. For example, my Thinkpad, which I bought just before the widescreen craze, has a 15" monitor with a native resolution of 1600 x 1200. The newer 15.4" Thinkpads run at most 1680 x 1050. So (if you do the math) you get fewer pixels in a wider (but not shorter) package. With desktop monitors, you pay a price in terms of desk space used. Two 1680 x 1050 monitors will simply take up more of your desk than two 1600 x 1200 monitors (assuming equal dot pitch). More scrolling If you compare a 1680 x 1050 monitor to a 1600 x 1200 monitor, you get 80 extra pixels of width but 150 fewer pixels of height. The height reduction means you lose approximately 11 lines of code. That's less you can see on the screen at one time and more scrolling you have to do. This harms productivity, maybe not dramatically, but insidiously. Less room for wide panels Widescreens also mean you lose space for wide but short panels common in programming environments. If you use Visual Studio, for example, your code window will be that much shorter when viewing the Find Results, Task List, or Error List (all of which I use frequently). This isn't to say the 80 pixels of extra width you get with widescreen would never be useful, but I tend to keep my lines of code short, so seeing more lines would be more valuable to me than seeing fewer, longer lines. What do you think? Do you agree/disagree? Are you now using one or more widescreen monitors for development? What resolution are you running on each? Do you ever miss the height of the traditional 4:3 monitor? Would you complain if your monitors were one inch narrower but two inches taller?

    Read the article

  • C++0x Overload on reference, versus sole pass-by-value + std::move?

    - by dean
    It seems the main advice concerning C++0x's rvalues is to add move constructors and move operators to your classes, until compilers default-implement them. But waiting is a losing strategy if you use VC10, because automatic generation probably won't be here until VC10 SP1, or in worst case, VC11. Likely, the wait for this will be measured in years. Here lies my problem. Writing all this duplicate code is not fun. And it's unpleasant to look at. But this is a burden well received, for those classes deemed slow. Not so for the hundreds, if not thousands, of smaller classes. ::sighs:: C++0x was supposed to let me write less code, not more! And then I had a thought. Shared by many, I would guess. Why not just pass everything by value? Won't std::move + copy elision make this nearly optimal? Example 1 - Typical Pre-0x constructor OurClass::OurClass(const SomeClass& obj) : obj(obj) {} SomeClass o; OurClass(o); // single copy OurClass(std::move(o)); // single copy OurClass(SomeClass()); // single copy Cons: A wasted copy for rvalues. Example 2 - Recommended C++0x? OurClass::OurClass(const SomeClass& obj) : obj(obj) {} OurClass::OurClass(SomeClass&& obj) : obj(std::move(obj)) {} SomeClass o; OurClass(o); // single copy OurClass(std::move(o)); // zero copies, one move OurClass(SomeClass()); // zero copies, one move Pros: Presumably the fastest. Cons: Lots of code! Example 3 - Pass-by-value + std::move OurClass::OurClass(SomeClass obj) : obj(std::move(obj)) {} SomeClass o; OurClass(o); // single copy, one move OurClass(std::move(o)); // zero copies, two moves OurClass(SomeClass()); // zero copies, one move Pros: No additional code. Cons: A wasted move in cases 1 & 2. Performance will suffer greatly if SomeClass has no move constructor. What do you think? Is this correct? Is the incurred move a generally acceptable loss when compared to the benefit of code reduction?

    Read the article

  • Performing full screen grab in windows

    - by Steven Lu
    I am working an idea that involves getting a full capture of the screen including windows and apps, analyzing it, and then drawing items back onto the screen, as an overlay. I want to learn image processing techniques and I could get lots of data to work with if I can directly access the Windows screen. I could use this to build automation tools the likes of which have never been seen before. More on that later. I have full screen capture working for the most part. HWND hwind = GetDesktopWindow(); HDC hdc = GetDC(hwind); int resx = GetSystemMetrics(SM_CXSCREEN); int resy = GetSystemMetrics(SM_CYSCREEN); int BitsPerPixel = GetDeviceCaps(hdc,BITSPIXEL); HDC hdc2 = CreateCompatibleDC(hdc); BITMAPINFO info; info.bmiHeader.biSize = sizeof(BITMAPINFOHEADER); info.bmiHeader.biWidth = resx; info.bmiHeader.biHeight = resy; info.bmiHeader.biPlanes = 1; info.bmiHeader.biBitCount = BitsPerPixel; info.bmiHeader.biCompression = BI_RGB; void *data; hbitmap = CreateDIBSection(hdc2,&info,DIB_RGB_COLORS,(void**)&data,0,0); SelectObject(hdc2,hbitmap); Once this is done, I can call this repeatedly: BitBlt(hdc2,0,0,resx,resy,hdc,0,0,SRCCOPY); The cleanup code (I have no idea if this is correct): DeleteObject(hbitmap); ReleaseDC(hwind,hdc); if (hdc2) { DeleteDC(hdc2); } Every time BitBlt is called it grabs the screen and saves it in memory I can access thru data. Performance is somewhat satisfactory. BitBlt executes in 50 milliseconds (sometimes as low as 33ms) at 1920x1200x32. What surprises me is that when I switch display mode to 16 bit, 1920x1200x16, either through my graphics settings beforehand, or by using ChangeDisplaySettings, I get a massively improved screen grab time between 1ms and 2ms, which cannot be explained by the factor of two reduction in bit-depth. Using CreateDIBSection (as above) offers a significant speed up when in 16-bit mode, compared to if I set up with CreateCompatibleBitmap (6-7ms/f). Does anybody know why dropping to 16bit causes such a speed increase? Is there any hope for me to grab 32bit at such speeds? if not for the color depth, but for not forcing a change of screen buffer modes and the awful flickering.

    Read the article

  • Why Fusion Middleware matters to Oracle Applications and Fusion Applications customers?

    - by Harish Gaur
    Did you miss this general session on Monday morning presented by Amit Zavery, VP of Oracle Fusion Middleware Product Management? There will be a recording made available shortly and in the meanwhile, here is a recap. Amit presented 5 strategies customers can leverage today to extend their applications. Figure 1: 5 Oracle Fusion Middleware strategies to extend Oracle Applications & Oracle Fusion Apps 1. Engage Everyone – Provide intuitive and social experience for application users using Oracle WebCenter 2. Extend Enterprise – Extend Oracle Applications to mobile devices using Oracle ADF Mobile 3. Orchestrate Processes – Automate key organization processes across on-premise & cloud applications using Oracle BPM Suite & Oracle SOA Suite 4. Secure the core – Provide single sign-on and self-service provisioning across multiple apps using Oracle Identity Management 5. Optimize Performance – Leverage Exalogic stack to consolidate multiple instance and improve performance of Oracle Applications Session included 3 demonstrations to illustrate these strategies. 1. First demo highlighted significance of mobile applications for unlocking existing investment in Applications such as EBS. Using a native iPhone application interacting with e-Business Suite, demo showed how expense approval can be mobile enabled with enhanced visibility using BI dashboards. 2. Second demo showed how you can extend a banking process in Siebel and Oracle Policy Automation with Oracle BPM Suite.Process starts in Siebel with a customer requesting a loan, and then jumps to OPA for loan recommendations and decision making and loan processing with approvals in handled in BPM Suite. Once approvals are completed Siebel is updated to complete the process. 3. Final demo showcased FMW components inside Fusion Applications, specifically WebCenter. Boeing, Underwriter Laboratories and Electronic Arts joined this quest and discussed 3 different approaches of leveraging Fusion Middleware stack to maximize their investment in Oracle Applications and/or Fusion Applications technology. Let’s briefly review what these customers shared during the session: 1. Extend Fusion Applications We know that Oracle Fusion Middleware is the underlying technology infrastructure for Oracle Fusion Applications. Architecturally, Oracle Fusion Apps leverages several components of Oracle Fusion Middleware from Oracle WebCenter for rich collaborative interface, Oracle SOA Suite & Oracle BPM Suite for orchestrating key underlying processes to Oracle BIEE for dash boarding and analytics. Boeing talked about how they are using Oracle BPM Suite 11g, a key component of Oracle Fusion Middleware with Oracle Fusion Apps to transform their supply chain. Tim Murnin, Director of Supply Chain talked about Boeing’s 5 year supply chain transformation journey. Boeing’s Integrated and Information Management division began with automation of critical RFQ process using Oracle BPM Suite. This 1st phase resulted in 38% reduction in labor costs for RFP. As a next step in this effort, Boeing is now creating a platform to enable electronic Order Management. Fusion Apps are playing a significant role in this phase. Boeing has gone live with Oracle Fusion Product Hub and efforts are underway with Oracle Fusion Distributed Order Orchestration (DOO). So, where does Oracle BPM Suite 11g fit in this equation? Let me explain. Business processes within Fusion Apps are designed using 2 standards: Business Process Execution Language (BPEL) and Business Process Modeling Notation (BPMN). These processes can be easily configured using declarative set of tools. Boeing leverages Oracle BPM Suite 11g (which supports BPMN 2.0) and Oracle SOA Suite (which supports BPEL) to “extend” these applications. Traditionally, customizations are done within an app using native technologies. But, instead of making process changes within Fusion Apps, Boeing has taken an approach of building “extensions” layer on top of the application. Fig 2: Boeing’s use of Oracle BPM Suite to orchestrate key supply chain processes across Fusion Apps 2. Maximize Oracle Applications investment Fusion Middleware appeals not only to Fusion Apps customers, but is also leveraged by Oracle E-Business Suite, PeopleSoft, Siebel and JD Edwards customers significantly. Using Oracle BPM Suite and Oracle SOA Suite is the recommended extension strategy for Oracle Fusion Apps and Oracle Applications Unlimited customers. Electronic Arts, E-Business Suite customer, spoke about their strategy to transform their order-to-cash process using Oracle SOA Suite, Oracle Foundation Packs and Oracle BAM. Udesh Naicker, Sr Director of IT at Elecronic Arts (EA), discussed how growth of social and digital gaming had started to put tremendous pressure on EA’s existing IT infrastructure. He discussed the challenge with millions of micro-transactions coming from several sources – Microsoft Xbox, Paypal, several service providers. EA found Order-2-Cash processes stretched to their limits. They lacked visibility into these transactions across the entire value chain. EA began by consolidating their E-Business Suite R11 instances into single E-Business Suite R12. EA needed to cater to a variety of service requirements, connectivity methods, file formats, and information latency. Their integration strategy was tactical, i.e., using file uploads, TIBCO, SQL scripts. After consolidating E-Business suite, EA standardized their integration approach with Oracle SOA Suite and Oracle AIA Foundation Pack. Oracle SOA Suite is the platform used to extend E-Business Suite R12 and standardize 60+ interfaces across several heterogeneous systems including PeopleSoft, Demantra, SF.com, Workday, and Managed EDI services spanning on-premise, hosted and cloud applications. EA believes that Oracle SOA Suite 11g based extension strategy has helped significantly in the followings ways: - It helped them keep customizations out of E-Business Suite, thereby keeping EBS R12 vanilla and upgrade safe - Developers are now proficient in technology which is also leveraged by Fusion Apps. This has helped them prepare for adoption of Fusion Apps in the future Fig 3: Using Oracle SOA Suite & Oracle e-Business Suite, Electronic Arts built new platform for order processing 3. Consolidate apps and improve scalability Exalogic is an optimal platform for customers to consolidate their application deployments and enhance performance. Underwriter Laboratories talked about their strategy to run their mission critical applications including e-Business Suite on Exalogic. Christian Anschuetz, CIO of Underwriter Laboratories (UL) shared how UL is on a growth path - $1B to $2.5B in 5 years- and planning a significant business transformation from a not-for-profit to a for-profit business. To support this growth, UL is planning to simplify its IT environment and the deployment complexity associated with ERP applications and technology it runs on. Their current applications were deployed on variety of hardware platforms and lacked comprehensive disaster recovery architecture. UL embarked on a mission to deploy E-Business Suite on Exalogic. UL’s solution is unique because it is one of the first to deploy a large number of Oracle applications and related Fusion Middleware technologies (SOA, BI, Analytical Applications AIA Foundation Pack and AIA EBS to Siebel UCM prebuilt integration) on the combined Exalogic and Exadata environment. UL is planning to move to a virtualized architecture toward the end of 2012 to securely host external facing applications like iStore Fig 4: Underwrites Labs deployed e-Business Suite on Exalogic to achieve performance gains Key takeaways are: - Fusion Middleware platform is certified with major Oracle Applications Unlimited offerings. Fusion Middleware is the underlying technological infrastructure for Fusion Apps - Customers choose Oracle Fusion Middleware to extend their applications (Apps Unlimited or Fusion Apps) to keep applications upgrade safe and prepare for Fusion Apps - Exalogic is an optimum platform to consolidate applications deployments and enhance performance

    Read the article

< Previous Page | 5 6 7 8 9 10 11  | Next Page >