Search Results

Search found 212 results on 9 pages for 'bothered'.

Page 6/9 | < Previous Page | 2 3 4 5 6 7 8 9  | Next Page >

  • download management

    - by Jonathan
    I download many files, usually 2 or 3 a day, often 10ish. Some of them are duplicates because I just can't be bothered to find the original in my downloads folder. I have previously tried DAP and used that to create a new subfolder for each day's download. yet I have found this insufficient as sometimes I wish to find files by name/file type or I have multiple parts of downloads over more than one day. Another problem I have found is zips/rars/etc after downloading them and extracting them I then have the zip and the folder. I like it like on a Mac where it automatically extracts the zip after it has been downloaded and removes the zip. What I'd like to be able to do is sort the downloads by date, but dynamically so they are just in the big downloads folder, but I can just press a button and it will show me all the files from a particular site, or from a particular day or by a certain file type. Is there any software that will do this? I use Chrome as a browser but also have Firefox and like that. Jonathan

    Read the article

  • What's needed in a complete ASP.NET environment?

    - by Christian W
    We have a ASP3.0 application with a few ASP.NET (2.0) dittys mixed in. (Our longtime goal is to migrate everything to ASP.NET but that's not important for this issue) Our current test/deploy workflow is like this: 1 Use notepad++ or VS2008 to fix a bug/feature (depending on what I have open) 2 Open my virtual test-server 3 Copy the fixed file over, either with explorer, or if I can be bothered to open it, WinMerge 4 Test that the fix works 5 Close the virtual test-server 6 Connect to our host with VPN 7 Use WinMerge to update the files necessary 8 Pray to higher powers that the production environment is not so different that something bombs. To make things worse, only I have access to my "test-server". So I'm the only one testing it. I really want to make this a bit more robust, I even have a subversion setup running. But I always forget to commit changes... And I don't even work in my checked out folder, but a copy of what is currently in production... Can someone recommend some good reading on deploying, testing, staging and stuff like that. I currently use VS2008 and want to use subversion or GIT (or any other free VCS). Since I'm the only developer, teamsystem is not really an option (cost-related). I have found myself developing an "improved" feature, only to find a bug in the same feature in the production system. And since my "improved" feature incorporated deleting some old functionality, I have to fix bugs directly in production... That's not a fun feeling... (I have inherited this system recently... So it's not directly my fault that it is like this ;) )

    Read the article

  • Looking for Remote Control that works with everything (even Windows 7 Media Center)

    - by T Reddy
    Using my Google-Fu, it seems that the most basic of things one gets with any DVR is the remote control. Had I known it would be difficult just to get a consumer IR receiver for Windows 7 I may not have bothered to build an HTPC. But too late, I already have the HTPC ready to go (minus the CETON card...) So I'm moving away from TiVo, I hate paying the monthly fees and my box is ancient. I'm looking for these solutions to my HTCP setup...I want to: Switch audio from HDMI to SPDIF via the remote control (i.e., switch from TV to Receiver) (as a side note, the built-in audio on the mobo has software to do this). Pressing the volume button on the remote will always change the TV's volume (or the Receiver's if possible) and NOT the PC's volume. The remote/receiver works well around 25 feet. Bonus if the IR Receiver can work with my existing TiVo remote (or other remotes laying around the house) I read a review of the Bluetooth TiVo remote...it sounds promising...but I'm not sure if it is great for Windows 7 HTPC?

    Read the article

  • How to make project auto-estimate duration based on work?

    - by Bruno Brant
    This one has bothered me for a long while. I like to do estimates thinking on how much time a certain task will take (I'm in TI business), so, let's say, it takes 12 hours to build a program. Now, let's say I tell Project that my beginning date is today. If I allocate one resource to this task, it means that the task will last 1,5 days, implying that it will end tomorrow. But right now, that is not what it's doing. I say that the task will take 1 hour, and when I add a resource to it, it allocate the resource at [13%] basis, which means that the duration is still fixed... project is trying to make the task last for a day. I have, on many occasions, accomplished this. What I do is build a plan based on these rough estimates for effort, then I allocate tasks to resources. Times conflict, so I level resources and then Project magically tells me how long, in days, will it take. But every time I have to start estimating again, I end up having trouble on how to make project work like that.

    Read the article

  • College network - can I point non-domain student computers to our SUS server?

    - by Joel Coel
    Since I started here 3 months ago, one of the things that's really bothered me about the way this network is setup is something that shows up on the daily bandwidth consumption report. I get a list of top-visited sites by hits and by size, and invariably the top site (to the point that it's bigger than all the other top sites combined) is au.download.windowsupdate.com. We're pulling in ~30GB/day in windows updates. This is every day, not just after a patch Tuesday. After a patch day, it jumps closer to 40GB for a couple days. The key here is that almost none if it is by machines that I'm responsible for. My machines are for the most part fully patched, and when they're not they'll pull from a SUS server, so new updates are downloaded only once. It used to be closer to 50GB/day because most of the machines in our computer labs use DeepFreeze and weren't applying updates correctly, but that's fixed now. So the problem is definitely student-owned machines in the dorms, some of which are re-downloading the same updates in background each day, over and over. I'd love to have these machines start pulling from our SUS server. Then, if they don't ever actually install them at least they're not leeching bandwidth from our public internet connection. Any ideas on how to resolve the situation?

    Read the article

  • Why is writing to my external hard drive slow, while benchmarks show fast writing?

    - by matix2267
    I have an iOmega eGo 320GB portable drive connected through USB2.0 to my laptop running Windows Vista. It's been working fine for quite some time until recently it became very slow when writing e.g. when copying ~300MB movie over to the drive at first it is extremely fast but it actually doesn't write it only puts in cache and then hangs on last 10-20MBs for about a minute. When copying larger files it's the same story: starts fast but then slows down to ~5MB/s (sometimes even slower down to 2MB/s). Strange thing is that I have always had caching disabled for this drive (it was disabled by default and I never bothered changing it). At first I thought that the disk is dying so I checked S.M.A.R.T. values and everything is fine there. I also run chkdsk and it seemed to fix the problem - it worked fast for a few minutes but then it slowed down again. I also tried plugging it into another USB port - no difference. Additionally I noticed that reading under certain circumstances is sometimes slower e.g. loading times for some games are ~10 times longer, whereas simple copying files from this drive to my internal HDD is fast. I ran a speed benchmark using CrystalDiskMark with a 5x100MB run and strangely got these results: read write (MB/s) Seq 33.05 28.25 512k 17.30 15.27 4k 0.267 0.372 4kQD32 0.510 0.260 This is different from what most other people have (I've found many threads about slow disk write while googling but all of them were slow on benchmarks too) which is why I decided to post this problem here. BTW most of the time when writing (or sometimes reading) the activity led is mostly idle (blinks a while and then stops for longer, sometimes has slower blinks ~1 sek, sometimes goes off for a few seconds - extremely long blink :) ) but when benchmarking, defragmenting or just reading (copying from this drive, installing apps from installers there, watching HD videos) it is blinking really fast (like it should) and there are no slowdowns. It shouldn't be driver issue unless stock Windows drivers have some issues I'm not aware of.

    Read the article

  • RouterLess, house-wired network using multiple powerline adapters

    - by Cliff Arnell
    related to the 'old days' of one ethernet cable tapped with Ts for each monitor.... my question might be very simple... or not. I have an over-the-air internet provider with a wire dish with a powered transceiver and cat5 cable out of the providers supplied modem. I'm presently connecting the output of the modem into my wireless router which sends the internet signal all over the house. Standard stuff, I believe. My Question. Can I just connect the output of the modem into 1 powerline adapter and tie all my equipment such as computer, printer, laptop, Tivo recorder, etc. into 1-each local powerline adapters located near each devices resulting in a 'house-wired' network and no router? I'm bothered by the idea that my over-the-air provider might be using something in my router to establish and keep my IP connection alive. I did have to configure the router for my IP, a router which, in my proposed scenario, would no longer exist. Thank you for your help.

    Read the article

  • ubuntu preseed installation keep missing mirror files

    - by JackWu
    Install ubuntu12.04.2 with preseed file, but there is one buggy problem about preseed mirror setting. The symptom here is installing process got stuck. So I track down the log file, and find out the real problem, the installation is looking for a file that's not there. This is just one of them, another pops up if I faked this file. This all happened during preseed, so I believe preseed has something to do with this. I google ubuntu preseed mirror and find this post saying: # If you select ftp, the mirror/country string does not need to be set. #d-i mirror/protocol string ftp d-i mirror/country string manual d-i mirror/http/hostname string archive.ubuntu.com d-i mirror/http/directory string /ubuntu d-i mirror/http/proxy string # Alternatively: by default, the installer uses CC.archive.ubuntu.com where # CC is the ISO-3166-2 code for the selected country. You can preseed this # so that it does so without asking. #d-i mirror/http/mirror select CC.archive.ubuntu.com # Suite to install. #d-i mirror/suite string lucid # Suite to use for loading installer components (optional). #d-i mirror/udeb/suite string lucid # Components to use for loading installer components (optional). #d-i mirror/udeb/components multiselect main, restricted I wonder the difference between d-i mirror/http/hostname and d-i mirror/http/mirror, I mean they all specify a mirror, right? In my preseed file, this is no d-i mirror/http/mirror, and d-i mirror/http/hostname points to my own repo as you might notice in the previous image. Here is my question: Does preseed fetches file/resource from internet, if I use local repo? Why it's looking for file that's not even there? This has bothered for quite time, many thanks in advance to anyone who might give any help.

    Read the article

  • Ubuntu 10.10 - disaster - what other linux for beginner?

    - by A-ha
    Guys, I've tried to install ubuntu (desktop and notebook ed) on my laptop and unortunately I have to say that as despite the fact that installation process supposed to be easy I couldn't finish installation of this system - didn't detect my keyboard or rather lost my keyboard as soon as I tried to switch on/off pad on my laptop. After I've discovered that, I started all over again (this time without touching my laptop's pad during installation) and yes, eventually it get to the end of installation. Unfortunately, when I've tried to switch my pad (sometimes I just do not want to use a mouse) the whole system froze. So I had to restart it with the power button and this time I didn't touch pad at all, plugged in mouse and tried to rearrange taskbars according to my liking (all taskbars on the top side of the screen and auto-hide on) and I gave up. It is so unfinished that I just can't be bothered to use it. I would like to have one linux system on my machine so I started googling and most of the links are to either ubuntu (which I just do not want to touch for now) and suse or commercial versions of linux. I do not really mind paying for something (and having experience with ubuntu I'd rather pay and have something pro then get it free and discover that it's unusable). So could someone please provide short list of linux distros which would be appropriate for a beginner, and I don't mind paying for it, I just want it to be a professional product.

    Read the article

  • 3 Monitor PCI-e Graphics card on Linux (without tremendous pain)?

    - by N Rahl
    As we are all painfully aware, the only way to get multiple monitors AND compositing (Compiz) on Linux is to use a single graphics card that can drive both (or in my case all three) screens. I bought a Radeon 5750 specifically because it claims to able to drive 3 monitors. I can plug in 3 monitors (2 DVI, 1 HDMI) and the Catalyst Control Center shows all 3, but only 2 can be enabled at a time. The exact message is: The current settings cannot be applied. Possible issues may include: - Display(s) cannot be enabled. - Setting(s) cannot be applied due to insufficient video memory. So I'm going to assume that either the 5750 doesn't support 3 monitors, OR, more likely, ATI couldn't be bothered to add that support to their Linux drivers. So this is a multipart question: First, can anyone suggest a PCI Express Graphics card that can run 3 screens on linux without tremendous pain? I'm looking for something where you install the driver and all three screens "just work". Does such a card exist? Second, if you have a 5750, have you been able to get it to do 3 monitors? I'm running Ubuntu 10.04 at the moment.

    Read the article

  • Can a OS be copied from one hard drive to another and still boot?

    - by AlexMorley-Finch
    Background My computer gets stuck on the make and model screen after the BIOS screen, aka the Toshiba screen. After some research I've realized that the problem is the hard drive. I'm using an old 250gb model that USED to be used for backup purposes, however I loaded windows 7 ultimate onto it This hard drive has trouble getting up to full RPM therefore cannot boot correctly until its warmed up. meaning that my pc needs to be restarted several times before it boots (once it took my 13 reboots to get my pc on!) From my research its either that, or lack of power supply, and I've tried multiple PSUs. Question I have my OS and all my files on this 250gb HDD... If I were to literally open the explorer, and copy EVERYTHING (including hidden files obviously) from this 250gb, to a spare 500gb I've got knocking about... Will it boot if I just copy everything? I cannot be bothered to load another OS onto my PC so if there is a way I can just copy the existing one over from one HDD to another and have it boot normally. This would be epic! I've heard about HDD cloning software. But before I purchase and/or download this software, I need to know if i can just copy the OS over through the windows explorer

    Read the article

  • Linux: 3 Monitor PCI-e Graphics card (without tremendous pain)?

    - by N Rahl
    As we are all painfully aware, the only way to get multiple monitors AND compositing (Compiz) on Linux is to use a single graphics card that can drive both (or in my case all three) screens. I bought a Radeon 5750 specifically because it claims to able to drive 3 monitors. I can plug in 3 monitors (2 DVI, 1 HDMI) and the Catalyst Control Center shows all 3, but only 2 can be enabled at a time. I'll post the exact error message here soon, but it's very useless. So I'm going to assume that either the 5750 doesn't support 3 monitors, OR, more likely, ATI couldn't be bothered to add that support to their Linux drivers. So this is a multipart question: First, can anyone suggest a PCI Express Graphics card that can run 3 screens on linux without tremendous pain? I'm looking for something where you install the driver and all three screens "just work". Does such a card exist? Second, if you have a 5750, have you been able to get it to do 3 monitors? I'm running Ubuntu 10.04 at the moment. Thanks, Nick

    Read the article

  • 3 Monitor PCI-e Graphics card (without tremendous pain)?

    - by N Rahl
    As we are all painfully aware, the only way to get multiple monitors AND compositing (Compiz) on Linux is to use a single graphics card that can drive both (or in my case all three) screens. I bought a Radeon 5750 specifically because it claims to able to drive 3 monitors. I can plug in 3 monitors (2 DVI, 1 HDMI) and the Catalyst Control Center shows all 3, but only 2 can be enabled at a time. The exact message is: The current settings cannot be applied. Possible issues may include: - Display(s) cannot be enabled. - Setting(s) cannot be applied due to insufficient video memory. So I'm going to assume that either the 5750 doesn't support 3 monitors, OR, more likely, ATI couldn't be bothered to add that support to their Linux drivers. So this is a multipart question: First, can anyone suggest a PCI Express Graphics card that can run 3 screens on linux without tremendous pain? I'm looking for something where you install the driver and all three screens "just work". Does such a card exist? Second, if you have a 5750, have you been able to get it to do 3 monitors? I'm running Ubuntu 10.04 at the moment.

    Read the article

  • GIT and Django Projects

    - by Garfonzo
    I have two servers, a Dev server and a Production server. The Production server runs a live Django site, while the Dev server has a copy of the Django project. I use the Dev server to work on the Django site, make improvements, fix bugs, etc. Once I am satisfied with how the Dev version is working, I move the whole Django directory from the Dev server and replace the same directory on the Production server. The two servers are not on the same LAN so the process is not straight forward. There are a few issues with this that I am having so far. Moving the whole directory is laborious and time consuming If I only change a few files, it is even move tedious to replace a few files than the whole directory since the project is getting fairly large and I worry that I'll miss something I often run into permission issues after I've moved things It's super inefficient, and, due to lack of time, I haven't bothered figuring out a new method. Now it's just getting out of hand and i need to address the situation. I am thinking I need to move to a GIT repository for this process. But my question is how would I set this all up? Do I host the repository on the Production server, pull from the Dev server, do work, then commit? Then I would pull from the Production server (same server the repo is hosted on) to run the current working version? Do I host the repo on the Dev Server, pulling from the same server to do work on the repo, then pull a working version onto the Production server? Should I be hosting the repo on a different server than the Production server and the Dev server (a third server)? Are there any special considerations with Django and repos that I need to worry about? Thanks for the help :)

    Read the article

  • cannot make ubuntu 64-bit v12.04 install work

    - by honestann
    I decided it was time to update my ubuntu (single boot) computer from 64-bit v10.04 to 64-bit v12.04. Unfortunately, for some reason (or reasons) I just can't make it work. Note that I am attempting a fresh install of 64-bit v12.04 onto a new 3TB hard disk, not an upgrade of the 1TB hard disk that contains my working 64-bit v10.04 installation. To perform the attempted install of v12.04 I unplug the SATA cable from the 1TB drive and plug it into the 3TB drive (to avoid risking damage to my working v10.04 installation). I downloaded the ubuntu 64-bit v12.04 install DVD ISO file (~1.6 GB) from the ubuntu releases webpage and burned it onto a DVD. I have downloaded the DVD ISO file 3 times and burned 3 of these installation DVDs (twice with v10.04 and once with my winxp64 system), but none of them work. I run the "check disk" on the DVDs at the beginning of the installation process to assure the DVD is valid. When installation completes and the system boots the 3TB drive, it reports "unknown filesystem". After installation on the 250GB drives, the system boots up fine. During every install I plug the same SATA cable (sda) into only one disk drive (the 3TB or one of the 250GB drives) and leave the other disk drives unconnected (for simplicity). It is my understanding that 64-bit ubuntu (and 64-bit linux in general) has no problem with 3TB disk drives. In the BIOS I have tried having EFI set to "enabled" and "auto" with no apparent difference (no success). I never bothered setting the BIOS to "non-EFI". I have tried partitioning the drive in a few ways to see if that makes a difference, but so far it has not mattered. Typically I manually create partitions something like this: 8GB /boot ext4 8GB swap 3TB / ext4 But I've also tried the following, just in case it matters: 8GB boot efi 8GB swap 8GB /boot ext4 3TB / ext4 Note: In the partition dialog I specify bootup on the same drive I am partitioning and installing ubuntu v12.04 onto. It is a VERY DANGEROUS FACT that the default for this always comes up with the wrong drive (some other drive, generally the external drive). Unless I'm stupid or misunderstanding something, this is very wrong and very dangerous default behavior. Note: If I connect the SATA cable to the 1TB drive that has been my ubuntu 64-bit v10.04 system drive for the past 2 years, it boots up and runs fine. I guess there must be a log file somewhere, and maybe it gives some hints as to what the problem is. I should be able to boot off the 1TB drive with the 3TB drive connected as a secondary (non-boot) drive and get the log file, assuming there is one and someone tells me the name (and where to find it if the name is very generic). After installation on the 3TB drive completes and the system reboots, the following prints out on a black screen: Loading Operating System ... Boot from CD/DVD : Boot from CD/DVD : error: unknown filesystem grub rescue> Note: I have two DVD burners in the system, hence the duplicate line above. Note: I install and boot 64-bit ubuntu v12.04 on both of my 250GB in this same system, but still cannot make the 3TB drive boot. Sigh. Any ideas? ========== motherboard == gigabyte 990FXA-UD7 CPU == AMD FX-8150 8-core bulldozer @ 3.6 GHz RAM == 8GB of DDR3 in 2 sticks (matched pair) HDD == seagate 3TB SATA3 @ 7200 rpm (new install 64-bit v12.04 FAILS) HDD == seagate 1TB SATA3 @ 7200 rpm (64-bit v10.04 WORKS for two years) HDD == seagate 250GB SATA2 @ 7200 rpm (new install 64-bit v12.04 WORKS) HDD == seagate 250GB SATA2 @ 7200 rpm (new install 64-bit v12.04 WORKS) GPU == nvidia GTX-285 ??? == no overclocking or other funky business USB == external seagate 2TB HDD for making backups DVD == one bluray burner (SATA) DVD == one DVD burner (SATA) 64-bit ubuntu v10.04 has booted and run fine on the seagate 1TB drive for 2 years.

    Read the article

  • Inequality joins, Asynchronous transformations and Lookups : SSIS

    - by jamiet
    It is pretty much accepted by SQL Server Integration Services (SSIS) developers that synchronous transformations are generally quicker than asynchronous transformations (for a description of synchronous and asynchronous transformations go read Asynchronous and synchronous data flow components). Notice I said “generally” and not “always”; there are circumstances where using asynchronous transformations can be beneficial and in this blog post I’ll demonstrate such a scenario, one that is pretty common when building data warehouses. Imagine I have a [Customer] dimension table that manages information about all of my customers as a slowly-changing dimension. If that is a type 2 slowly changing dimension then you will likely have multiple rows per customer in that table. Furthermore you might also have datetime fields that indicate the effective time period of each member record. Here is such a table that contains data for four dimension members {Terry, Max, Henry, Horace}: Notice that we have multiple records per customer and that the [SCDStartDate] of a record is equivalent to the [SCDEndDate] of the record that preceded it (if there was one). (Note that I am on record as saying I am not a fan of this technique of storing an [SCDEndDate] but for the purposes of clarity I have included it here.) Anyway, the idea here is that we will have some incoming data containing [CustomerName] & [EffectiveDate] and we need to use those values to lookup [Customer].[CustomerId]. The logic will be: Lookup a [CustomerId] WHERE [CustomerName]=[CustomerName] AND [SCDStartDate] <= [EffectiveDate] AND [EffectiveDate] <= [SCDEndDate] The conventional approach to this would be to use a full cached lookup but that isn’t an option here because we are using inequality conditions. The obvious next step then is to use a non-cached lookup which enables us to change the SQL statement to use inequality operators: Let’s take a look at the dataflow: Notice these are all synchronous components. This approach works just fine however it does have the limitation that it has to issue a SQL statement against your lookup set for every row thus we can expect the execution time of our dataflow to increase linearly in line with the number of rows in our dataflow; that’s not good. OK, that’s the obvious method. Let’s now look at a different way of achieving this using an asynchronous Merge Join transform coupled with a Conditional Split. I’ve shown it post-execution so that I can include the row counts which help to illustrate what is going on here: Notice that there are more rows output from our Merge Join component than on the input. That is because we are joining on [CustomerName] and, as we know, we have multiple records per [CustomerName] in our lookup set. Notice also that there are two asynchronous components in here (the Sort and the Merge Join). I have embedded a video below that compares the execution times for each of these two methods. The video is just over 8minutes long. View on Vimeo  For those that can’t be bothered watching the video I’ll tell you the results here. The dataflow that used the Lookup transform took 36 seconds whereas the dataflow that used the Merge Join took less than two seconds. An illustration in case it is needed: Pretty conclusive proof that in some scenarios it may be quicker to use an asynchronous component than a synchronous one. Your mileage may of course vary. The scenario outlined here is analogous to performance tuning procedural SQL that uses cursors. It is common to eliminate cursors by converting them to set-based operations and that is effectively what we have done here. Our non-cached lookup is performing a discrete operation for every single row of data, exactly like a cursor does. By eliminating this cursor-in-disguise we have dramatically sped up our dataflow. I hope all of that proves useful. You can download the package that I demonstrated in the video from my SkyDrive at http://cid-550f681dad532637.skydrive.live.com/self.aspx/Public/BlogShare/20100514/20100514%20Lookups%20and%20Merge%20Joins.zip Comments are welcome as always. @Jamiet Share this post: email it! | bookmark it! | digg it! | reddit! | kick it! | live it!

    Read the article

  • Release Management as Orchestra

    - by ericajanine
    I read an excellent, concise article (http://www.buildmeister.com/articles/software_release_management_best_practices) on the basics of release management practices. In the article, it states "Release Management is often likened to the conductor of an orchestra, with the individual changes to be implemented the various instruments within it." I played in music ensembles for years, so this is especially close to my heart as example. I learned most of my discipline from hours and hours of practice at the hand of a very skilled conductor and leader. I also learned that the true magic in symphonic performance is one where everyone involved is focused on one sound, one goal. In turn, that solid focus creates a sound and experience bigger than just mechanics alone accomplish. In symphony, a conductor's true purpose is to make you, a performer, better so the overall sound and end product is better. The big picture (the performance of the composition) is the end-game, and all musicians in the orchestra know without question their part makes up an important but incomplete piece of that performance. A good conductor works with each section (e.g. group) to ensure their individual pieces are solid. Let's restate: The conductor leads and is responsible for ensuring those pieces are solid. While the performers themselves are doing the work, the conductor is the final authority on when the pieces are ready or not. If not, the conductor initiates the efforts to get them ready or makes the decision to scrap their parts altogether for the sake of an overall performance. Let it sink in, because it's clear--It is not the performer's call if they play their part as agreed, it's the conductor's final call to allow it. In comparison, if a software release manager is a conductor, the only way for that manager to be effective is to drive the overarching process and execution of individual pieces of a software development lifecycle. It does not mean the release manager performs each and every piece, it means the release manager has oversight and influence because the end-game is a successful software enhancin a useable environment. It means the release manager, not the developer or development manager, has the final call if something goes into a software release. Of course, this is not a process of autocracy or dictation of absolute rule, it's cooperative effort. But the release manager must have the final authority to make a decision if something is ready to be added to the bigger piece, the overall symphony of software changes being considered for package and release. It also goes without saying a release manager, like a conductor, must have full autonomy and isolation from other software groups. A conductor is the one on the podium waving a little stick at the each section and cueing them for their parts, not yelling from the back of the room while also playing a tuba and taking direction from the horn section. I have personally seen where release managers are relegated to being considered little more than coordinators, red-tapers to "satisfy" the demands of an audit group without being bothered to actually respect all that a release manager gives a group willing to employ them fully. In this dysfunctional scenario, development managers, project managers, business users, and other stakeholders have been given nearly full clearance to demand and push their agendas forward, causing a tail-wagging-the-dog scenario where an inherent conflict will ensue. Depending on the strength, determination for peace, and willingness to overlook a built-in expectation that is wrong, the release manager here must face the crafted conflict head-on and diffuse it as quickly as possible. Then, the release manager must clearly make a case why a change cannot be released without negative impact to all parties involved. If a political agenda is solely driving a software release, there IS no symphony, there is no "software lifecycle". It's just out-of-tune noise. More importantly, there is no real conductor. Sometimes, just wanting to make a beautiful sound is not enough. If you are a release manager, are you freed up enough to move, to conduct the sections of software creation to ensure a solid release performance is possible? If not, it's time to take stock in what your role actually is and see if that is what you truly want to achieve in your position. If you are, then you can successfully build your career and that of the people in your groups to create truly beautiful software (music) together.

    Read the article

  • ArchBeat Link-o-Rama Top 10 for August 2012

    - by Bob Rhubart
    The Top 10 most popular items shared via the OTN ArchBeat Facebook page for the month of August 2012. Now Available: Oracle SQL Developer 3.2 (3.2.09.23) New features include APEX listener, UI enhancements, and 12c database support. The Role of Oracle VM Server for SPARC in a Virtualization Strategy In this article, Matthias Pfutzner discusses hardware, desktop, and operating system virtualization, along with various Oracle virtualization technologies, including Oracle VM Server for SPARC. How to Manually Install Flash Player Plugin to see the Oracle Enterprise Manager Performance Page | Kai Yu So, you're a DBA and you want to check the Performance page in Oracle Enterprise Manager (11g or 12c). So you click the Performance tab and… nothing. Zip. Nada. The Flash plugin is a no-show. Relax! Oracle ACE Director Kai Yu shows you what you need to do to see all the pretty colors instead of that dull grey screen. Relationally Challenged (CX - CRM - EQ/RQ/CRQ) | Chris Warticki Self-proclaimed Oracle Support "spokesmodel" Chris Chris Warticki has some advice for those interested in Customer Relationship Management: "How about we just dumb it down, strip it to the core, keep it simple and LISTEN?! No more focus groups, no more surveys, and no need to gather more data. We have plenty of that. Why not just provide the customer what they are asking for?" Free WebLogic Server Course | Middleware Magic So you want to sharpen your Oracle WebLogic Server skills, but you prefer to skip the whole classroom bit and don't want to be bothered with dealing with an instructor? No problem! Oracle ACE Rene van Wijk, a prolific Middleware Magic blogger, has information on an Oracle WebLogic course you can take on your own time, at your own pace. Oracle VM VirtualBox 4.1.20 released Oracle VM VirtualBox 4.1.20 was just released at the community and Oracle download sites, reports the Fat Bloke. This is a maintenance release containing bug fixes and stability improvements. Optimizing OLTP Oracle Database Performance using Dell Express Flash PCIe SSDs | Kai Yu Oracle ACE Director Kai Yu shares resources based on "several extensive performance studies on a single node Oracle 11g R2 database as well as a two node 11gR2 Oracle Real Application clusters (RAC) database running on Dell PowerEdge R720 servers with Dell Express Flash PCIe SSDs on Oracle Enterprise Linux 6.2 platform." Oracle ACE sessions at Oracle OpenWorld With so many great sessions at this year's event, building your Oracle OpenWorld schedule can involve making a lot of tough choices. But you'll find that the sessions led by Oracle ACEs just might be the icing on the cake for your OpenWorld experience. MySQL Update: The Cleveland MySQL Meetup (Independence, OH) Oracle MySQL team member Benjamin Wood, a MySQL engineer and five year veteran of the MySQL organization, will speak at the Cleveland MySQL Meetup event on September 12. The presentation will include a MySQL 5.5 Overview, Oracle's Roadmap for MySQL, including specifics on MySQL 5.6, best practices and how to overcome development and operational MySQL challenges, and the new MySQL commercial extensions. Click the link for time and location information. Parsing XML in Oracle Database | Martijn van der Kamp Martijn van der Kamp's post deals with processing XML in PL/SQL code and processing the data into the database. Thought for the Day "Walking on water and developing software from a specification are easy if both are frozen." — Edward V. Berard Source: SoftwareQuotes.com

    Read the article

  • How do you educate your teammates without seeming condescending or superior?

    - by Dan Tao
    I work with three other guys; I'll call them Adam, Brian, and Chris. Adam and Brian are bright guys. Give them a problem; they will figure out a way to solve it. When it comes to OOP, though, they know very little about it and aren't particularly interested in learning. Pure procedural code is their MO. Chris, on the other hand, is an OOP guy all the way -- and a cocky, condescending one at that. He is constantly criticizing the work Adam and Brian do and talking to me as if I must share his disdain for the two of them. When I say that Adam and Brian aren't interested in learning about OOP, I suspect Chris is the primary reason. This hasn't bothered me too much for the most part, but there have been times when, looking at some code Adam or Brian wrote, it has pained me to think about how a problem could have been solved so simply using inheritance or some other OOP concept instead of the unmaintainable mess of 1,000 lines of code that ended up being written instead. And now that the company is starting a rather ambitious new project, with Adam assigned to the task of getting the core functionality in place, I fear the result. Really, I just want to help these guys out. But I know that if I come across as just another holier-than-thou developer like Chris, it's going to be massively counterproductive. I've considered: Team code reviews -- everybody reviews everybody's code. This way no one person is really in a position to look down on anyone else; besides, I know I could learn plenty from the other members on the team as well. But this would be time-consuming, and with such a small team, I have trouble picturing it gaining much traction as a team practice. Periodic e-mails to the team -- this would entail me sending out an e-mail every now and then discussing some concept that, based on my observation, at least one team member would benefit from learning about. The downside to this approach is I do think it could easily make me come across as a self-appointed expert. Keeping a blog -- I already do this, actually; but so far my blog has been more about esoteric little programming tidbits than straightforward practical advice. And anyway, I suspect it would get old pretty fast if I were constantly telling my coworkers, "Hey guys, remember to check out my new blog post!" This question doesn't need to be specifically about OOP or any particular programming paradigm or technology. I just want to know: how have you found success in teaching new concepts to your coworkers without seeming like a condescending know-it-all? It's pretty clear to me there isn't going to be a sure-fire answer, but any helpful advice (including methods that have worked as well as those that have proved ineffective or even backfired) would be greatly appreciated. UPDATE: I am not the Team Lead on this team. Chris is. UPDATE 2: Made community wiki to accord with the general sentiment of the community (fancy that).

    Read the article

  • SQL Sentry Truth-Telling and Disk Configuration

    - by AjarnMark
    Recently, SQL Sentry told me something about my SQL Server disk configurations that I just didn’t want to believe, but alas, it was true. Several days ago I posted my First Impressions of the SQL Sentry Power Suite.  Today’s post could fall into the category of, “Hey, as long as you have that fancy tool…”  Unfortunately, it also falls into the category of an overloaded worker taking someone else’s word for the truth, not verifying it with independent fact-checking, and then making decisions based on that.  Here’s my story… I’m not exactly an Accidental DBA (or Involuntary DBA as Paul Randal calls it).  I came to this company five years ago as a lead application developer with extensive experience in database design and development.  I worked my way into management, and along the way, took over the DBA responsibilities.  Fortunately, our systems run pretty smoothly most of the time, but I’m always looking for ways to make them better and to fit into my understanding of best practices.  When I took over as DBA, I inherited a SQL 2000 server with about 30 databases on it supporting our main systems, and a SQL 2005 server with multiple instances.  Both of these servers were configured with the Operating System and Application files on the C drive, data files on a different drive letter, and log files on a third drive letter.  Even before I took over as DBA, I verified that this was true with a previous server administrator, and that these represented actual separate disks.  He stated that they did, and I thought that all was well. Then one day, I’m poking around inside the SQL Sentry Performance Advisor, checking out features as I am evaluating whether to purchase the product, and I come across a Disk Configuration section.  The first thing I notice is that the drives do not have the proper partition offset, which was not at all surprising to me given the age of the installation and the relative newness of that topic.  But what threw me for a loop was that the graphic display appeared to be telling me that I did not in fact have three separate drives (or arrays) but rather had two, and that the log files were merely on a separate volume on the same physical array as the OS.  I figured that I must be reading it wrong so I scanned the Help file, but that just seemed to confirm my interpretation.  Then I thought, “there must be something wrong with the demo version of the software!  This can’t be right!”  But just to double-check, I went to our current server admin to talk it over with him, and sure enough, SQL Sentry was telling the truth! I was stunned!  I quickly went through the grieving process…denial…anger…reconciliation.  Here was something that I thought was such a basic truth that was turned upside down.  OK, granted, this wasn’t disastrous.  Our databases didn’t suddenly grind to a halt.  I didn’t get calls late at night inquiring about the sudden downturn in performance.  But it was a bit of a shock to the system, in a good way, to jolt me out of taking what I had believed as the truth for granted, and instead to Trust, but Verify! Yes, before someone else points it out, I know that there are”free” disk management tools built-in to Windows that would have told me the same thing if I had only looked at them; I did not have to buy a fancy tool to tell me that, but the fact is, until I was evaluating the tool, I had just gone with what I was told, and never bothered to check what was actually there. So, what things do you believe to be true but you actually never verified?

    Read the article

  • Should I break contract early?

    - by cbang
    About 7 months ago I made the switch from a 5 year permie role (as a support developer in C#) to a contract role. I did this because I was stagnating in my old role. The extra cash contracting is really helping too. Unfortunately my team leader has taken a dislike to me from day 1. He regularly tells me I went out contracting too early, and frequently remarks that people in their 20's have no idea what they are talking about (I am 29). I was recently given the task of configuring our reports via our in house reporting library. It works off of a database driven criteria base, with controls being loaded as needed. The configs can get fairly complex, with controls having various levels of dependency on each other. I had a short time frame to get 50 reports working, and I was told to just get the basic configuration done, after which they will be handed over to the reporting team for fine tuning, then the test team. Our updated system was deployed 2 weeks ago, and it turned out that about 15 reports had issues causing incorrect data to be returned. Upon investigation I discovered that the reporting team hadn't even looked at them, and the test team hadn't bothered to test the reports. In spite of this, my team leader has told me that it is 100% my fault. As a result, our help desk got hit hard. I worked back until 2am that night to fix the highest priority issues (on my wedding anniversary!). The next day I arrive at work at 7:45 am to continue with the fixes. I got no thanks, but keep getting repeatedly told by my manager that "I fucked up" and "this is all my fault". I told my team leader I would spend part of my weekend working to fix the remaining issues. His response was "so you fucking should! you fucked it all up!" in front of the rest of the team. I responded "No worries." and left. I spent a decent chunk of my weekend working on it. Within 2 business days of finding out about the issues, I had all the medium and high priority issues fixed. The only comments my team leader has made to me in the last 2 weeks is to tell me how I have caused a big mess, and to tell me it was all my fault. I get this multiple times a day. If I make any jokes to anyone else in the team, I get told not to be a smartass... even though the rest of the team jokes throughout the day. Apart from that, all I get is angry looks any time I am anywhere near the guy. I don't give any response other than "alright" or silence when he starts giving me a hard time. Today we found out that the pilot release for the next stage has been pushed back. My team leader has said this was caused by me (but the higher ups said no such thing). He also said I have "no understanding of the ramifications of my actions". My question is, should I break contract (I am contracted until June 30) and find another role? No one else in my team will speak up in my favour, as they are contractors too and have no interest in rocking the boat. I could complain to my team leaders boss, but I can't see that helping, as I will still be stuck in the same team. As this is my first contract, I imagine getting the next one will be hard without a reference. I can't figure out if this guy is trying to get me fired up to provoke a confrontation (the guy loves conflict), or if he is just venting anger, or what. Copping this blame day after day is really wearing me down and making me depressed... especially since I have a wife and kid to support).

    Read the article

  • Green (Screen) Computing

    - by onefloridacoder
    I recently was given an assignment to create a UX where a user could use the up and down arrow keys, as well as the tab and enter keys to move through a Silverlight datagrid that is going be used as part of a high throughput data entry UI. And to be honest, I’ve not trapped key codes since I coded JavaScript a few years ago.  Although the frameworks I’m using made it easy, it wasn’t without some trial and error.    The other thing that bothered me was that the customer tossed this into the use case as they were delivering the use case.  Fine.  I’ll take a whack at anything and beat up myself and beg (I’m not beyond begging for help) the community for help to get something done if I have to. It wasn’t as bad as I thought and I thought I would hopefully save someone a few keystrokes if you wanted to build a green screen for your customer.   Here’s the ValueConverter to handle changing the strings to decimals and then back again.  The value is a nullable valuetype so there are few extra steps to take.  Usually the “ConvertBack()” method doesn’t get addressed but in this case we have two-way binding and the converter needs to ensure that if the user doesn’t enter a value it will remain null when the value is reapplied to the model object’s setter.  1: using System; 2: using System.Windows.Data; 3: using System.Globalization; 4:  5: public class NullableDecimalToStringConverter : IValueConverter 6: { 7: public object Convert(object value, Type targetType, object parameter, System.Globalization.CultureInfo culture) 8: { 9: if (!(((decimal?)value).HasValue)) 10: { 11: return (decimal?)null; 12: } 13: if (!(value is decimal)) 14: { 15: throw new ArgumentException("The value must be of type decimal"); 16: } 17:  18: NumberFormatInfo nfi = culture.NumberFormat; 19: nfi.NumberDecimalDigits = 4; 20:  21: return ((decimal)value).ToString("N", nfi); 22: } 23:  24: public object ConvertBack(object value, Type targetType, object parameter, System.Globalization.CultureInfo culture) 25: { 26: decimal nullableDecimal; 27: decimal.TryParse(value.ToString(), out nullableDecimal); 28:  29: return nullableDecimal == 0 ? null : nullableDecimal.ToString(); 30: } 31: }            The ConvertBack() method uses TryParse to create a value from the incoming string so if the parse fails, we get a null value back, which is what we would expect.  But while I was testing I realized that if the user types something like “2..4” instead of “2.4”, TryParse will fail and still return a null.  The user is getting “puuu-lenty” of eye-candy to ensure they know how many values are affected in this particular view. Here’s the XAML code.   This is the simple part, we just have a DataGrid with one column here that’s bound to the the appropriate ViewModel property with the Converter referenced as well. 1: <data:DataGridTextColumn 2: Header="On-Hand" 3: Binding="{Binding Quantity, 4: Mode=TwoWay, 5: Converter={StaticResource DecimalToStringConverter}}" 6: IsReadOnly="False" /> Nothing too magical here.  Just some XAML to hook things up.   Here’s the code behind that’s handling the DataGridKeyup event.  These are wired to a local/private method but could be converted to something the ViewModel could use, but I just need to get this working for now. 1: // Wire up happens in the constructor 2: this.PicDataGrid.KeyUp += (s, e) => this.HandleKeyUp(e);   1: // DataGrid.BeginEdit fires when DataGrid.KeyUp fires. 2: private void HandleKeyUp(KeyEventArgs args) 3: { 4: if (args.Key == Key.Down || 5: args.Key == Key.Up || 6: args.Key == Key.Tab || 7: args.Key == Key.Enter ) 8: { 9: this.PicDataGrid.BeginEdit(); 10: } 11: }   And that’s it.  The ValueConverter was the biggest problem starting out because I was using an existing converter that didn’t take nullable value types into account.   Once the converter was passing back the appropriate value (null, “#.####”) the grid cell(s) and the model objects started working as I needed them to. HTH.

    Read the article

  • Things I've noticed with DVCS

    - by Wes McClure
    Things I encourage: Frequent local commits This way you don't have to be bothered by changes others are making to the central repository while working on a handful of related tasks.  It's a good idea to try to work on one task at a time and commit all changes at partitioned stopping points.  A local commit doesn't have to build, just FYI, so a stopping point doesn't mean a build point nor a point that you can push centrally.  There should be several of these in any given day.  2 hours is a good indicator that you might not be leveraging the power of frequent local commits.  Once you have verified a set of changes works, save them away, otherwise run the risk of introducing bugs into it when working on the next task.  The notion of a task By task I mean a related set of changes that can be completed in a few hours or less.  In the same token don’t make your tasks so small that critically related changes aren’t grouped together.  Use your intuition and the rest of these principles and I think you will find what is comfortable for you. Partial commits Sometimes one task explodes or unknowingly encompasses other tasks, at this point, try to get to a stopping point on part of the work you are doing and commit it so you can get that out of the way to focus on the remainder.  This will often entail committing part of the work and continuing on the rest. Outstanding changes as a guide If you don't commit often it might mean you are not leveraging your version control history to help guide your work.  It's a great way to see what has changed and might be causing problems.  The longer you wait, the more that has changed and the harder it is to test/debug what your changes are doing! This is a reason why I am so picky about my VCS tools on the client side and why I talk a lot about the quality of a diff tool and the ability to integrate that with a simple view of everything that has changed.  This is why I love using TortoiseHg and SmartGit: they show changed files, a diff (or two way diff with SmartGit) of the current selected file and a commit message all in one window that I keep maximized on one monitor at all times. Throw away / stash commits There is extreme value in being able to throw away a commit (or stash it) that is getting out of hand.  If you do not commit often you will have to isolate the work you want to commit from the work you want to throw away, which is wasted productivity and highly prone to errors.  I find myself doing this about once a week, especially when doing exploratory re-factoring.  It's much easier if I can just revert all outstanding changes. Sync with the central repository daily The rest of us depend on your changes.  Don't let them sit on your computer longer than they have to.  Waiting increases the chances of merge conflict which just decreases productivity.  It also prohibits us from doing deploys when people say they are done but have not merged centrally.  This should be done daily!  Find a way to partition the work you are doing so that you can sync at least once daily. Things I discourage: Lots of partial commits right at the end of a series of changes If you notice lots of partial commits at the end of a set of changes, it's likely because you weren't frequently committing, nor were you watching for the size of the task expanding beyond a single commit.  Chances are this cost you productivity if you use your outstanding changes as a guide, since you would have an ever growing list of changes. Committing single files Committing single files means you waited too long and no longer understand all the changes involved.  It may mean there were overlapping changes in single files that cannot be isolated.  In either case, go back to the suggestions above to avoid this.  Committing frequently does not mean committing frequently right at the end of a day's work. It should be spaced out over the course of several tasks, not all at the end in a 5 minute window.

    Read the article

  • Company wants to write custom project management tool, rather then use third party product.

    - by Jason Evans
    At the company I work, we are really wanting to get into the agile methodology for developing software. One thing that I'm not excited about is the fact that management wants us to build a custom project management feature inside the company's Intranet. I think this is a total waste of time. There are many great third party tools available (e.g. Axosoft OnTime) that can do everything we need, and more. For how much development time it would cost us to build our own project management module, we could buy numerous licences for a third party product. One concern is that, whilst we are writing code for a client, and using our custom Intranet project management module, we find bugs in the module that need fixing ASAP. That means having to stop work on the client code to fix the Intranet. That just puts shivers down my spine. Another worry I have is lack of functionality. This custom module is going to be so basic, that it will just feel really crap to use. That might sound a bit snooty, but for goodness sake, many third party tools are so feature rich, that the idea of having to write our own tool makes feel very uneasy. In fact, I can't be bothered. What do you guys think? I'm going to raise this issue with my boss, since I feel it's such an important topic to talk about. EDIT: Thanks for the great responses, much appreciated. To summarize some of them: Money Naturally my boss does want to save money, by not forking out a few hundred £'s for licences. However, for us to write a custom tool, it will take x number of days, multiplied by approx £500, which is our costs. I don't see the business value in this. Management have mentioned that they want to sell the Intranet as a product in the future, but it's so custom to our needs (and downright basic), that in order to give it to another client, I can see us having to fork a version of the code and rebuild the majority of it anyway. So it's not like we're gaining anything there in reuse. Features Having our own custom module means not feature bloat - only the functionality we require will be in the product. My issue is that there are plenty of free, open-source project management tools out there with minimal features already. So even if cost is an issue, we could look into open-source. Again it all boils down to the fact that I don't see the point in writing a project management tool in this day and age. It's a bit like writing your own web browser - why?, what's the point? Although management are asking for this tool, just because they are, it does not mean I'm going to please them and do it just because they asked for it. If something does not make sense, then I will raise it as a concern. At the end of the day, it's the developers who write the code, it's the developers who make money for a business. Thus, as far I'm concerned, the devs have a very big role in deciding how a company should manage projects and what tools are used. "I am Spartan, argh!" :) Hmm, I've not been able to make this question a wiki for some reason, thus I'm going to have to pick an answer to accept. Cheers. Jas.

    Read the article

  • Issues glVertexAttribPointer last 2 parameters?

    - by NoobScratcher
    Introduction Hello I will start out by explaining my setup, showing samples as I go along explaining the situation. I'm using these tools: OpenGL 3.3 GLSL 330 C++ Problem The problem is when I render the wavefront obj 3d model it gives a very weird visual glitch the model was supposed to be a square but instead its a triangluated mess with parts of the vertexes pointing in a stretched direction in massive amounts towards the bottom left side of the frustum.... Explanation: I'm using std::vectors to store my wavefront .obj model data using sscanf to get the floating point values into the structure members x,y,z and store them into the Points structure variable p; int index = IndexAssigner(1, 1); ifstream file (list[index].c_str() ); points.push_back(Point()); Point p; int face[4]; while (!file.eof() ) { char modelbuffer[10000]; file.getline(modelbuffer, 10000); switch(modelbuffer[0]) { case 'v' : sscanf(modelbuffer, "v %f %f %f", &p.x, &p.y, &p.z); points.push_back(p); break; case 'f': sscanf(modelbuffer, "f %d %d %d %d", face, face+1, face+2, face+3 ); faces.push_back(face[0]); faces.push_back(face[1]); faces.push_back(face[2]); faces.push_back(face[3]); } //Turn on FileReader aka "RENDER CODE" FileReader = true; } then I render the Points vector using the .data() member of std::vectors to the frustum. Other declarations: int numfloats = 4; float* point=reinterpret_cast<float*>(&points[0]); int num_bytes=numfloats*sizeof(float); Vector declarations: struct Point {float x, y , z; }; std::vector<int>faces; std::vector<Point>points; Render code: glGenBuffers(1, &vertexbuffer); glGenTextures(1, &ModelTexture); glBindBuffer(GL_ARRAY_BUFFER, vertexbuffer); glBindTexture(GL_TEXTURE_3D, ModelTexture); glTexImage2D(GL_TEXTURE_2D, 0,GL_RGBA, ModelSurface->w, ModelSurface->h, 0, GL_BGR, GL_UNSIGNED_BYTE, ModelSurface->pixels); glTexParameteri(GL_TEXTURE_2D, GL_TEXTURE_MIN_FILTER, GL_NEAREST); glTexParameteri(GL_TEXTURE_2D, GL_TEXTURE_MAG_FILTER, GL_NEAREST); glBufferData(GL_ARRAY_BUFFER, sizeof(points), points.data(), GL_STATIC_DRAW); glVertexAttribPointer(3, 3, GL_FLOAT, GL_FALSE,num_bytes ,points.data()); glEnableVertexAttribArray(3); //Translation Process GLfloat TranslationMatrix[] = { 1.0, 0.0, 0.0, 0.0, 0.0, 1.0, 0.0, 0.0, 0.0, 0.0, 1.0, 1.0, 0.0, 0.0, 0.0, 1.0 }; //Send Translation Matrix up to the vertex shader glUniformMatrix4fv(translation, 1, TRUE, TranslationMatrix); glDrawElements( GL_QUADS, faces.size(), GL_UNSIGNED_INT, faces.data()); I tried looking at what was causing this and went through every function every parameter ,etc looked at the man pages. Then found out that it could be my glVertexAttribPointer. Here are the man pages for glVertexAttribPointer http://www.opengl.org/sdk/docs/man/xhtml/glVertexAttribPointer.xml The last 2 parameters is my problem How do I write those 2 last parameters do I try putting the data from Points into it?. glVertexAttribPointer(3, 3, GL_FLOAT, GL_FALSE,num_bytes ,points.data()); How does it work with vectors? Is it fast?* if you can not be bothered too look at the man pages here is the scripts coming from the man pages directly. Stride Specifies the byte offset between consecutive generic vertex attributes. If stride is 0, the generic vertex attributes are understood to be tightly packed in the array. The initial value is 0. Pointer Specifies a pointer to the first component of the first generic vertex attribute in the array. The initial value is 0. If you want my full source - http://ideone.com/fPfkg Thanks Again if you do read this.

    Read the article

< Previous Page | 2 3 4 5 6 7 8 9  | Next Page >