Search Results

Search found 3722 results on 149 pages for 'choice'.

Page 17/149 | < Previous Page | 13 14 15 16 17 18 19 20 21 22 23 24  | Next Page >

  • Is the “jQuery programming style” a kind of Reactive programming?

    - by Peter Krauss
    jQuery is a Javascript library and framework, but when we are programming with jQuery into DOM problems/solutions, we can practice a style quite different of programming... We can read about jQuery at Wikipedia, The set of jQuery core features — DOM element selections, traversal and manipulation —, enabled by its selector engine (...), created a new "programming style", fusing algorithms and DOM-data-structures This question is similar to the "subquestion-3" of this question but not so generic. The focus here is about this new kind of "programming style"... So, the question: Is the "jQuery programming style in DOM context" a new paradign? Or it is more one example of reactive programming (not "cell-oriented" but "DOM-node oriented") or another one? We have no "standard taxonomy of paradigms", so, please, in your answer, indicate also your "best choice for Wikipedia Paradign". Example: if you understand that "jQuery programming DOM" is like "awk filtering data", your choice can be event-driven.

    Read the article

  • Windows Azure Use Case: Infrastructure Limits

    - by BuckWoody
    This is one in a series of posts on when and where to use a distributed architecture design in your organization's computing needs. You can find the main post here: http://blogs.msdn.com/b/buckwoody/archive/2011/01/18/windows-azure-and-sql-azure-use-cases.aspx  Description: Physical hardware components take up room, use electricity, create heat and therefore need cooling, and require wiring and special storage units. all of these requirements cost money to rent at a data-center or to build out at a local facility. In some cases, this can be a catalyst for evaluating options to remove this infrastructure requirement entirely by moving to a distributed computing environment. Implementation: There are three main options for moving to a distributed computing environment. Infrastructure as a Service (IaaS) The first option is simply to virtualize the current hardware and move the VM’s to a provider. You can do this with Microsoft’s Hyper-V product or other software, build the systems and host them locally on fewer physical machines. This is a good option for canned-applications (where you have to type setup.exe) but not as useful for custom applications, as you still have to license and patch those servers, and there are hard limits on the VM sizes. Software as a Service (SaaS) If there is already software available that does what you need, it may make sense to simply purchase not only the software license but the use of it on the vendor’s servers. Microsoft’s Exchange Online is an example of simply using an offering from a vendor on their servers. If you do not need a great deal of customization, have no interest in owning or extending the source code, and need to implement a solution quickly, this is a good choice. Platform as a Service (PaaS) If you do need to write software for your environment, your next choice is a Platform as a Service such as Windows Azure. In this case you no longer manager physical or even virtual servers. You start at the code and data level of control and responsibility, and your focus is more on the design and maintenance of the application itself. In this case you own the source code and can extend or change it as you see fit. An interesting side-benefit to using Windows Azure as a PaaS is that the Application Fabric component allows a hybrid approach, which gives you a basis to allow on-premise applications to leverage distributed computing paradigms. No one solution fits every situation. It’s common to see organizations pick a mixture of on-premise, IaaS, SaaS and PaaS components. In fact, that’s a great advantage to this form of computing - choice. References: 5 Enterprise steps for adopting a Platform as a Service: http://blogs.msdn.com/b/davidmcg/archive/2010/12/02/5-enterprise-steps-for-adopting-a-platform-as-a-service.aspx?wa=wsignin1.0  Application Patterns for the Cloud: http://blogs.msdn.com/b/kashif/archive/2010/08/07/application-patterns-for-the-cloud.aspx

    Read the article

  • SEOs: mobile version using AJAX: how to be properly read by SEOs?

    - by Olivier Pons
    Before anything else, I'd like to emphasize that I've already read this and this. Here's what I can do: Choice (1): create classical Web version with all products in that page - http://www.myweb.com. create mobile Web version with all products in the page and use jQuery Mobile to format all nicely. But this may be long to (load + format), and may provide bad user experience - http://m.myweb.com. Choice (2): create classical Web version with all products in that page create mobile Web version with almost nothing but a Web page showing "wait", then download all products in the page using AJAX and use jQuery Mobile to format all nicely. Showing a "wait, loading" message gives far more time to do whatever I want and may provide better user experience - http://m.myweb.com. Question: if I choose solution (2), google won't read anything on the mobile version (because all products will be downloaded in the page using AJAX), so it wont be properly read by SEOs. What / how shall I do?

    Read the article

  • SEOs: mobile version using AJAX: how to be properly read by crawlers?

    - by Olivier Pons
    Before anything else, I'd like to emphasize that I've already read this and this. Here's what I can do: First choice: create classical web version with all products in that page - http://www.myweb.com. create mobile web version with all products in the page and use jQuery Mobile to format all nicely but this may be long to (load + format), and may provide bad user experience - http://m.myweb.com. Second choice: create classical web version with all products in that page create mobile web version with almost nothing but a Web page showing wait, then download all products in the page using AJAX and use jQuery Mobile to format all nicely. Showing a wait, loading message gives far more time to do whatever I want and may provide better user experience - http://m.myweb.com. Question: if I choose the second solution, Google won't read anything on the mobile version (because all products will be downloaded in the page using AJAX), so it wont be properly read by crawlers. What / how shall I do?

    Read the article

  • SharePoint: Problem with BaseFieldControl

    - by Anoop
    Hi All, In below code in a Gird First column is BaseFieldControl from a column of type Choice of SPList. Secound column is a text box control with textchange event. Both the controls are created at rowdatabound event of gridview. Now the problem is that when Steps: 1) select any of the value from BaseFieldControl(DropDownList) which is rendered from Choice Column of SPList 2) enter any thing in textbox in another column of grid. 3) textchanged event fires up and in textchange event rebound the grid. Problem: the selected value becomes the first item or the default value(if any). but if i do not rebound the grid at text changed event it works fine. Please suggest what to do. using System; using Microsoft.SharePoint; using Microsoft.SharePoint.WebControls; using System.Web.UI; using System.Web.UI.WebControls; using System.Data; namespace SharePointProjectTest.Layouts.SharePointProjectTest { public partial class TestBFC : LayoutsPageBase { GridView grid = null; protected void Page_Load(object sender, EventArgs e) { try { grid = new GridView(); grid.ShowFooter = true; grid.ShowHeader = true; grid.AutoGenerateColumns = true; grid.ID = "grdView"; grid.RowDataBound += new GridViewRowEventHandler(grid_RowDataBound); grid.Width = Unit.Pixel(900); MasterPage holder = (MasterPage)Page.Controls[0]; holder.FindControl("PlaceHolderMain").Controls.Add(grid); DataTable ds = new DataTable(); ds.Columns.Add("Choice"); //ds.Columns.Add("person"); ds.Columns.Add("Curr"); for (int i = 0; i < 3; i++) { DataRow dr = ds.NewRow(); ds.Rows.Add(dr); } grid.DataSource = ds; grid.DataBind(); } catch (Exception ex) { } } void tx_TextChanged(object sender, EventArgs e) { DataTable ds = new DataTable(); ds.Columns.Add("Choice"); ds.Columns.Add("Curr"); for (int i = 0; i < 3; i++) { DataRow dr = ds.NewRow(); ds.Rows.Add(dr); } grid.DataSource = ds; grid.DataBind(); } void grid_RowDataBound(object sender, GridViewRowEventArgs e) { if (e.Row.RowType == DataControlRowType.DataRow) { SPWeb web = SPContext.Current.Web; SPList list = web.Lists["Source for test"]; SPField field = list.Fields["Choice"]; SPListItem item=list.Items.Add(); BaseFieldControl control = (BaseFieldControl)GetSharePointControls(field, list, item, SPControlMode.New); if (control != null) { e.Row.Cells[0].Controls.Add(control); } TextBox tx = new TextBox(); tx.AutoPostBack = true; tx.ID = "Curr"; tx.TextChanged += new EventHandler(tx_TextChanged); e.Row.Cells[1].Controls.Add(tx); } } public static Control GetSharePointControls(SPField field, SPList list, SPListItem item, SPControlMode mode) { if (field == null || field.FieldRenderingControl == null || field.Hidden) return null; try { BaseFieldControl webControl = field.FieldRenderingControl; webControl.ListId = list.ID; webControl.ItemId = item.ID; webControl.FieldName = field.Title; webControl.ID = "id_" + field.InternalName; webControl.ControlMode = mode; webControl.EnableViewState = true; return webControl; } catch (Exception ex) { return null; } } } }

    Read the article

  • Building small Ubuntu server - What hardware is recommended?

    - by 10robinho
    There are many of us who need to build small Ubuntu server. Problem is that in some countries it is hard to find and quite expensive to buy server motherboards and processors. And when one is building small server with limited budget, buying some Xenons is not really an option. So, are there any general recommendations for hardware (I think that motherboards are the main issue) that is stable and fast under Linux? I read that Intel should be the best choice for cpu + mbo combo. So, I was looking around for some Intel motherboards + i7 Ivy Bridge (like Intel DZ77BH-55K with Z77 chipset and Intel i7 3770K) but I've read that they have some issues with kernel, booting and USB ports. That is why I ask community if you have any experience with this. Maybe Intel is not the best choice here? Maybe ASUS or Gigabyte or _other company_ are more stable with Linux? I hope that this Q&As can help people in building stable Ubuntu server.

    Read the article

  • Is there any one standard framework for developing Python GUI apps.?

    - by RPK
    There are so many frameworks for writing GUI application using Python. But is there any one key standard framework? For example we have a bundle of .NET/C# on Visual Studio. I am thinking in other perspectives also. In future if I give an interview for a Python programmer job, which GUI framework will be considered? I also wonder, there is no IDE that integrates the GUI and Python language. Choice of flavor is good but over-choice becomes a distraction.

    Read the article

  • C++ - Constructor or Initialize Method to Startup

    - by Bob Fincheimer
    I want to determine when to do non-trivial initialization of a class. I see two times to do initialization: constructor and other method. I want to figure out when to use each. Choice 1: Constructor does initialization MyClass::MyClass(Data const& data) : m_data() { // does non-trivial initialization here } MyClass::~MyClass() { // cleans up here } Choice 2: Defer initialization to an initialize method MyClass::MyClass() : m_data() {} MyClass::Initialize(Data const& data) { // does non-trivial initialization here } MyClass::~MyClass() { // cleans up here } So to try and remove any subjectivity I want to figure out which is better in a couple of situations: Class that encapsulates a resource (window/font/some sort of handle) Class that composites resources to do something (a control/domain object) Data structure classes (tree/list/etc.) [Anything else you can think of] Things to analyze: Performance Ease of use by other developers How error-prone/opportunities for bugs [Anything else you can think of]

    Read the article

  • Should all foreign table references use foreign key constraints

    - by TecBrat
    Closely related to: Foreign key restrictions -> yes or no? I asked a question on SO and it led me to ask this here. If I'm faced with a choice of having a circular reference or just not enforcing the restraint, which is the better choice? In my particular case I have customers and addresses. I want an address to have a reference to a customer and I want each customer to have a default billing address id and a default shipping address id. I might query for all addresses that have a certain customer ID or I might query for the address with the ID that matches the default shipping or billing address ids. I'm not sure yet how the constraints (or lack of) will effect the system as my application and it's data age.

    Read the article

  • Diving into Scala with Cay Horstmann

    - by Janice J. Heiss
    A new interview with Java Champion Cay Horstmann, now up on otn/java, titled  "Diving into Scala: A Conversation with Java Champion Cay Horstmann," explores Horstmann's ideas about Scala as reflected in his much lauded new book,  Scala for the Impatient.  None other than Martin Odersky, the inventor of Scala, called it "a joy to read" and the "best introduction to Scala". Odersky was so enthused by the book that he asked Horstmann if the first section could be made available as a free download on the Typesafe Website, something Horstmann graciously assented to. Horstmann acknowledges that some aspects of Scala are very complex, but he encourages developers to simply stay away from those parts of the language. He points to several ways Java developers can benefit from Scala: "For example," he says, " you can write classes with less boilerplate, file and XML handling is more concise, and you can replace tedious loops over collections with more elegant constructs. Typically, programmers at this level report that they write about half the number of lines of code in Scala that they would in Java, and that's nothing to sneeze at. Another entry point can be if you want to use a Scala-based framework such as Akka or Play; you can use these with Java, but the Scala API is more enjoyable. " Horstmann observes that developers can do fine with Scala without grasping the theory behind it. He argues that most of us learn best through examples and not through trying to comprehend abstract theories. He also believes that Scala is the most attractive choice for developers who want to move beyond Java and C++.  When asked about other choices, he comments: "Clojure is pretty nice, but I found its Lisp syntax a bit off-putting, and it seems very focused on software transactional memory, which isn't all that useful to me. And it's not statically typed. I wanted to like Groovy, but it really bothers me that the semantics seems under-defined and in flux. And it's not statically typed. Yes, there is Groovy++, but that's in even sketchier shape. There are a couple of contenders such as Kotlin and Ceylon, but so far they aren't real. So, if you want to do work with a statically typed language on the JVM that exists today, Scala is simply the pragmatic choice. It's a good thing that it's such a nice choice." Learn more about Scala by going to the interview here.

    Read the article

  • JAXB: Unmarshalling does not always populate certain classes?

    - by user278458
    Hello, I have a JAXB class generation problem I was hoping to get some help with. Here's the part of the XML that is the source of my problem... Code: <xs:complexType name="IDType"> <xs:choice minOccurs="0" maxOccurs="2"> <xs:element name="DriversLicense" minOccurs="0" maxOccurs="1" type="an..35" /> <xs:element name="SSN" minOccurs="0" maxOccurs="1" type="an..35" /> <xs:element name="CompanyID" minOccurs="0" maxOccurs="1" type="an..35" /> </xs:choice> </xs:complexType> <xs:simpleType name="an..35"> <xs:restriction base="an"> <xs:maxLength value="35" /> </xs:restriction> </xs:simpleType> <xs:simpleType name="an"> <xs:restriction base="xs:string"> <xs:pattern value="[ !-~]*" /> </xs:restriction> </xs:simpleType> ...now this will generate JAXBElement types due the the "choice" with a "maxOccurs 1" . I want to avoid those, so I did that by modifying the code to use a "Wrapper" element and move the maxOccurs up to a sequence tag as follows... Code: <xs:complexType name="IDType"> <xs:sequence maxOccurs="2"> <xs:element name=Wrapper> <xs:complexType> <xs:choice> <xs:element name="DriversLicense" minOccurs="0" maxOccurs="1" type="an..35" /> <xs:element name="SSN" minOccurs="0" maxOccurs="1" type="an..35" /> <xs:element name="CompanyID" minOccurs="0" maxOccurs="1" type="an..35" /> </xs:choice> </xs:complexType> </xs:element> </xs:sequence> </xs:complexType> <xs:simpleType name="an..35"> <xs:restriction base="an"> <xs:maxLength value="35" /> </xs:restriction> </xs:simpleType> <xs:simpleType name="an"> <xs:restriction base="xs:string"> <xs:pattern value="[ !-~]*" /> </xs:restriction> </xs:simpleType> For class generating, looks like it works great - the JAXB element is replaced with a list of wrappers as String (i.e. List ) and compiles fine. However, when I unmarshall the actual XML data into the generated classes the data in the wrapper class is not populated - yet JAXB does not throw an exception. My question is: Do I need to change the schema a different way to make this work? Or is there something I can add/change/delete to the generated code or annotations? Appreciate any help you can offer! Thanks.

    Read the article

  • When should I use AtomPub?

    - by Gary Rowe
    I have been conducting some research into RESTful web service design and I've reached what I think is a key decision point so I thought I'd offer it up to the community to get some advice. In keeping with the principles of a RESTful architecture I want to present a discoverable API, so I will be supporting the various HTTP verbs as fully as possible. My difficulty comes with the choice of representation of those resources. You see, it would be easy for me to come up with my own API that covers how search results are to be presented and how links to other resources are provided, but this would be unique to my application. I've read about the Atom Publishing Protocol (RFC 5023), and how OData promotes its use, but it seems to add an extra level of abstraction over what is (currently) a rather simple API. So my question is, when should a developer select AtomPub as their choice of representation - if at all? And if not, what is the current recommended approach?

    Read the article

  • What does ENDOFMENU do?

    - by Raymond
    I have been given an example program, I am wondering what exactly the <<ENDOFMENU and ENDOFMENU does, won't it work the same if you leave it out and just use the while loop? #!/bin/sh echo "Menu test program..."; stop=0; while test $stop -eq 0; do cat<<ENDOFMENU 1: print the date 2,3 : print the current working directory 4: exit ENDOFMENU echo; echo -e "your choice?\c" read reply echo case $reply in "1") date ;; "2"|"3") pwd ;; "4") stop=1 ;; *) echo illegal choice esac done

    Read the article

  • Of transactions and Mongo

    - by Nuri Halperin
    Originally posted on: http://geekswithblogs.net/nuri/archive/2014/05/20/of-transactions-and-mongo-again.aspxWhat's the first thing you hear about NoSQL databases? That they lose your data? That there's no transactions? No joins? No hope for "real" applications? Well, you *should* be wondering whether a certain of database is the right one for your job. But if you do so, you should be wondering that about "traditional" databases as well! In the spirit of exploration let's take a look at a common challenge: You are a bank. You have customers with accounts. Customer A wants to pay B. You want to allow that only if A can cover the amount being transferred. Let's looks at the problem without any context of any database engine in mind. What would you do? How would you ensure that the amount transfer is done "properly"? Would you prevent a "transaction" from taking place unless A can cover the amount? There are several options: Prevent any change to A's account while the transfer is taking place. That boils down to locking. Apply the change, and allow A's balance to go below zero. Charge person A some interest on the negative balance. Not friendly, but certainly a choice. Don't do either. Options 1 and 2 are difficult to attain in the NoSQL world. Mongo won't save you headaches here either. Option 3 looks a bit harsh. But here's where this can go: ledger. See, and account doesn't need to be represented by a single row in a table of all accounts with only the current balance on it. More often than not, accounting systems use ledgers. And entries in ledgers - as it turns out – don't actually get updated. Once a ledger entry is written, it is not removed or altered. A transaction is represented by an entry in the ledger stating and amount withdrawn from A's account and an entry in the ledger stating an addition of said amount to B's account. For sake of space-saving, that entry in the ledger can happen using one entry. Think {Timestamp, FromAccountId, ToAccountId, Amount}. The implication of the original question – "how do you enforce non-negative balance rule" then boils down to: Insert entry in ledger Run validation of recent entries Insert reverse entry to roll back transaction if validation failed. What is validation? Sum up the transactions that A's account has (all deposits and debits), and ensure the balance is positive. For sake of efficiency, one can roll up transactions and "close the book" on transactions with a pseudo entry stating balance as of midnight or something. This lets you avoid doing math on the fly on too many transactions. You simply run from the latest "approved balance" marker to date. But that's an optimization, and premature optimizations are the root of (some? most?) evil.. Back to some nagging questions though: "But mongo is only eventually consistent!" Well, yes, kind of. It's not actually true that Mongo has not transactions. It would be more descriptive to say that Mongo's transaction scope is a single document in a single collection. A write to a Mongo document happens completely or not at all. So although it is true that you can't update more than one documents "at the same time" under a "transaction" umbrella as an atomic update, it is NOT true that there' is no isolation. So a competition between two concurrent updates is completely coherent and the writes will be serialized. They will not scribble on the same document at the same time. In our case - in choosing a ledger approach - we're not even trying to "update" a document, we're simply adding a document to a collection. So there goes the "no transaction" issue. Now let's turn our attention to consistency. What you should know about mongo is that at any given moment, only on member of a replica set is writable. This means that the writable instance in a set of replicated instances always has "the truth". There could be a replication lag such that a reader going to one of the replicas still sees "old" state of a collection or document. But in our ledger case, things fall nicely into place: Run your validation against the writable instance. It is guaranteed to have a ledger either with (after) or without (before) the ledger entry got written. No funky states. Again, the ledger writing *adds* a document, so there's no inconsistent document state to be had either way. Next, we might worry about data loss. Here, mongo offers several write-concerns. Write-concern in Mongo is a mode that marshals how uptight you want the db engine to be about actually persisting a document write to disk before it reports to the application that it is "done". The most volatile, is to say you don't care. In that case, mongo would just accept your write command and say back "thanks" with no guarantee of persistence. If the server loses power at the wrong moment, it may have said "ok" but actually no written the data to disk. That's kind of bad. Don't do that with data you care about. It may be good for votes on a pole regarding how cute a furry animal is, but not so good for business. There are several other write-concerns varying from flushing the write to the disk of the writable instance, flushing to disk on several members of the replica set, a majority of the replica set or all of the members of a replica set. The former choice is the quickest, as no network coordination is required besides the main writable instance. The others impose extra network and time cost. Depending on your tolerance for latency and read-lag, you will face a choice of what works for you. It's really important to understand that no data loss occurs once a document is flushed to an instance. The record is on disk at that point. From that point on, backup strategies and disaster recovery are your worry, not loss of power to the writable machine. This scenario is not different from a relational database at that point. Where does this leave us? Oh, yes. Eventual consistency. By now, we ensured that the "source of truth" instance has the correct data, persisted and coherent. But because of lag, the app may have gone to the writable instance, performed the update and then gone to a replica and looked at the ledger there before the transaction replicated. Here are 2 options to deal with this. Similar to write concerns, mongo support read preferences. An app may choose to read only from the writable instance. This is not an awesome choice to make for every ready, because it just burdens the one instance, and doesn't make use of the other read-only servers. But this choice can be made on a query by query basis. So for the app that our person A is using, we can have person A issue the transfer command to B, and then if that same app is going to immediately as "are we there yet?" we'll query that same writable instance. But B and anyone else in the world can just chill and read from the read-only instance. They have no basis to expect that the ledger has just been written to. So as far as they know, the transaction hasn't happened until they see it appear later. We can further relax the demand by creating application UI that reacts to a write command with "thank you, we will post it shortly" instead of "thank you, we just did everything and here's the new balance". This is a very powerful thing. UI design for highly scalable systems can't insist that the all databases be locked just to paint an "all done" on screen. People understand. They were trained by many online businesses already that your placing of an order does not mean that your product is already outside your door waiting (yes, I know, large retailers are working on it... but were' not there yet). The second thing we can do, is add some artificial delay to a transaction's visibility on the ledger. The way that works is simply adding some logic such that the query against the ledger never nets a transaction for customers newer than say 15 minutes and who's validation flag is not set. This buys us time 2 ways: Replication can catch up to all instances by then, and validation rules can run and determine if this transaction should be "negated" with a compensating transaction. In case we do need to "roll back" the transaction, the backend system can place the timestamp of the compensating transaction at the exact same time or 1ms after the original one. Effectively, once A or B visits their ledger, both transactions would be visible and the overall balance "as of now" would reflect no change.  The 2 transactions (attempted/ reverted) would be visible , since we do actually account for the attempt. Hold on a second. There's a hole in the story: what if several transfers from A to some accounts are registered, and 2 independent validators attempt to compute the balance concurrently? Is there a chance that both would conclude non-sufficient-funds even though rolling back transaction 100 would free up enough for transaction 117 (some random later transaction)? Yes. there is that chance. But the integrity of the business rule is not compromised, since the prime rule is don't dispense money you don't have. To minimize or eliminate this scenario, we can also assign a single validation process per origin account. This may seem non-scalable, but it can easily be done as a "sharded" distribution. Say we have 11 validation threads (or processing nodes etc.). We divide the account number space such that each validator is exclusively responsible for a certain range of account numbers. Sounds cunningly similar to Mongo's sharding strategy, doesn't it? Each validator then works in isolation. More capacity needed? Chop the account space into more chunks. So where  are we now with the nagging questions? "No joins": Huh? What are those for? "No transactions": You mean no cross-collection and no cross-document transactions? Granted - but don't always need them either. "No hope for real applications": well... There are more issues and edge cases to slog through, I'm sure. But hopefully this gives you some ideas of how to solve common problems without distributed locking and relational databases. But then again, you can choose relational databases if they suit your problem.

    Read the article

  • Have you tried to switch to different kind of language and find a job in a "new language" ?

    - by IAdapter
    I'm a Java programmer(J2EE/JEE), but I'm thinking about switching to C#. Does any of you have been in my position and have switched from Java to C# or C# to Java or C++ to Java, etc. ?? I'm NOT asking about switching between the same kind of languages, for example Java to Groovy/Scala/JRuby, C++ to C, VB to C#, C# to IronRuby/F#/VB.NET. Or if you company was C++, but has moved to Java(you had no choice and I'm about to make a choice). Side question: How hard was it to get a job in a "new language"?

    Read the article

  • Are C or C++ The Only Viable Languages for a GC

    - by user95312
    Background I have just finished writing a compiler for a functional language compiling to the JVM as a learning project. However, since I'm just doing this to learn, I thought it might be interesting to write a native backend and a RTS for it. As I've been planning out what this new backend will look like, the one point I'm stumbling on is the garbage collector. I've implemented the compiler in Haskell. But I have no desire to write the GC in Haskell since, while it may be possible, it'd suck. Question I've looked at several FOSS garbage collectors prior to posting and most of them were implemented in good old ANSI C. Is this still the most accepted choice for writing a GC nowadays? I've seen that this site tends to frown upon questions with multiple answers so I hope this will make it more specific: If some startup was writing a professional grade gc today, are the only viable choice for them C or C++? It's my first question here so please comment and let me know if this question is ill-suited for for programmers.

    Read the article

  • Website creation preparation [closed]

    - by Loki
    I am in the pre-coding phase of creating a website. I know that it will be account based (users have to register/login to use the features). I also know that the server will have to do certain operations that are timer based, that is to say that user will have events that will trigger at a point chosen by the user and do something. I am searching for a good choice in server-side technology, and was wondering what my options are and what the best choice is. I would prefer open technology and something that doesn't use interpreted languages (Java, .net). My first thought is PHP + PGSQL for serverside and HTML+CSS+JS for clients, but I am still looking at my options.

    Read the article

  • Mallocing an unsigned char array to store ints

    - by Max Desmond
    I keep getting a segmentation fault when i test the following code. I am currently unable to find an answer after having searched the web. a = (byte *)malloc(sizeof(byte) * x ) ; for( i = 0 ; i < x-1 ; i++ ) { scanf("%d", &y ) ; a[i] = y ; } Both y and x are initialized. X is the size of the array determined by the user. The segmentation fault is on the second to last integer to be added, i found this by adding printf("roar") ; before setting a[i] to y and entering one number at a time. Byte is a typedef of an unsigned char. Note: I've also tried using a[i] = (byte)y ; A is ininitalized as follows byte *a ; If you need to view the entire code it is this: #include <stdio.h> #include <stdlib.h> #include "sort.h" int p_cmp_f () ; int main( int argc, char *argv[] ) { int x, y, i, choice ; byte *a ; while( choice !=2 ) { printf( "Would you like to sort integers?\n1. Yes\n2. No\n" ) ; scanf("%d", &choice ) ; switch(choice) { case 1: printf( "Enter the length of the array: " ) ; scanf( "%d", &x ) ; a = (byte *)malloc(sizeof( byte ) * x ) ; printf( "Enter %d integers to add to the array: ", x ) ; for( i = 0 ; i < x -1 ; i++ ) { scanf( "%d", &y ) ; a[i] = y ; } switch( choice ) { case 1: bubble_sort( a, x, sizeof(int), p_cmp_f ) ; for( i = 0 ; i < x ; i++ ) printf( "%d", a[i] ; break ; case 2: selection_sort( a, x, sizeof(int), p_cmp_f ) ; for( i = 0 ; i < x; i++ ) printf( "%d", a[i] ; break ; case 3: insertion_sort( a, x, sizeof(int), p_cmp_f ) ; for( i = 0 ; i < x ; i++ ) printf( "%d", a[i] ; break ; case 4: merge_sort( a, x, sizeof(int), p_cmp_f ) ; for( i = 0 ; i < x ; i++ ) printf( "%d", a[i] ; break ; case 5: quick_sort( a, x, sizeof(int), p_cmp_f ) ; for( i = 0 ; i < x ; i++ ) printf( "%d", a[i] ; break ; default: printf("Enter either 1,2,3,4, or 5" ) ; break ; } case 2: printf( "Thank you for using this program\n" ) ; return 0 ; break ; default: printf( "Enter either 1 or 2: " ) ; break ; } } free(a) ; return 0 ; } int p_cmp_f( byte *element1, byte *element2 ) { return *((int *)element1) - *((int *)element2) ; }

    Read the article

  • A look at an example of anti-spam algorithm

    - by pragmaticCamel
    What is a good approach to an anti-spam algorithm for a website similar to reddit? Their anti-spam algorithm seems awfully broken (banning on words in the title and doing a horrible job for that matter). Considering a post spam because it has the word 'spam' in the title is really not a wise choice. Anyway, how can one approach such problem ? Are there any tools that help in such cases? Also, what are the /technical/ reasons behind reddit's choice not using reCAPTCHA on every post submission? It seems like a much better solution than what they have right now. Since reddit is basically a community-driven website why not give such power to the communities' trusted members?

    Read the article

  • In setting up dual Boot with Windows XP and Ubuntu, which OS do I install first?

    - by markl
    I'd like to install both Ubuntu 12.04 and Windows XP on a Dell laptop, and I was thinking about using a dual boot structure, and using the bulk of my hard drive as empty hard drive space to share files between the two operating systems (so choice of file system type is very important in this set-up). The kind of partitioning structure I would like to use is Partition 1 - Ubuntu 12.04 (root) (20GB) Partition 2 - Ubuntu /home (20GB) Partition 3 : Free Space (560GB) Partition 4 : Windows XP (35GB) Partition 5 : SWAP (3GB) (Total Hardrive Capacity is ~640GB) My question is; what is the best way to go about setting up this kind this system? Should I install Windows XP first and setup the partitions, and then install Ubuntu which I believe will install the GRUB bootloader for OS booting choice or Do I install Ubuntu first, setting up the available partitions and then perform a WIndows install? Please let me know if there is anything in this setup that I have left out and should know about, including things related to setting particular partitions as logical or primary, and whether the boot partition and the filesystem partition should actually be two separate partitions.

    Read the article

< Previous Page | 13 14 15 16 17 18 19 20 21 22 23 24  | Next Page >