Tech Off Thread

20 posts

Forum Read Only

This forum has been made read only by the site admins. No new threads or comments can be added.

Would you like to see a Visual Studio Beta 3?

Back to Forum: Tech Off
  • User profile image
    Sabot

    It seems that all is not well out there in VS2005 Beta 2 land. Clint Stotesbery has called for a vote on the MSDN Product Feedback Center for a Beta 3.

    In his problem statement he has written ... 

    There are still way too many bugs and performance issues. Too many issues get resolved as Postponed. A lot of changes are happenning in the CTPs. I'm guessing not many people are using the CTPs compared to beta 2. Most developers don't have time to play around with a CTP that doesn't have a Go Live license. I'd rather have a good product 6 months from now than a mediocre one in 3 months.”

    Anyway if you agree you can vote here.

  • User profile image
    irascian

    Good to see you back here sabot!

    This has already been discussed (although not very intelligently) in a thread I started yesterday
    http://channel9.msdn.com/ShowPost.aspx?PostID=99525 

  • User profile image
    Cryption

    MS is going to release VS05 in 05 no matter what.

  • User profile image
    JChung2006

    They've already announced the release date (November, 2005).  Considering that's three months from now, releasing a Beta 3 between now and then hardly seems productive.

    I understand where you're coming from though.  VS Beta 2 is very unstable.  I have no idea how stable the CTP releases since Beta 2 have been, since I am not an MSDN subscriber.  However, given the attitudes of the team handling bug reports, the only thing I can hope for at this point is that VS 2005 Service Pack 1 is scheduled for an early 2006 release.  They've really gotten bad with the "don't try hard to reproduce a user-submitted bug and close the bug rather than acknowledge the bug and fix it" approach.

    Because of this attitude, VS 2005 is not going to be a high-quality release, and what's worse, I don't think the people in charge of it realize it yet, because all they seem to care about are "ZBB's," which are highly inaccurate since most of their bugs are being classified as "not reproducible" instead of "fixed."

  • User profile image
    Tensor

    To be honest, I dont think we need a beta 3. I'm not sure what it would achieve right now other than push back the release date.

  • User profile image
    Sampy

    Let me jump in here.

    A lot of bugs get resolved "No Repro" because we try them on the latest bits and they don't repro. My team entered the RTM milestone with a list of known bugs and we kept finding them after beta 2 shipped. My features have undergone a lot of revision and polish since Beta 2 shipped so bugs people are reporting against Beta 2 are most likely fixed by now.

    Of course, we take your feedback very seriously. Just today in War room I saw a customer feedback bug that was so important it was classified as a recall class bug (if we shipped it, we'd do a QFE for it). Trust me, we're doing our best to make sure we ship quality.

    Do I think we could do better? YES! My team and many others around DevDiv are working on ways to make this process better so we introduce less bugs and fix them faster. Hopefully you'll see this pay off in Orcas.

  • User profile image
    Larsenal

    How about when I copy 'n paste a bunch of the same control, provide some means of tabbing through particular properties as I set them.

    For example, I am collecting visitor information (name, address, etc...) so I drag a bunch of TextBoxes onto the design surface.  I'd like to be able to select all of those text boxes with (CTRL), then have the property editor treat them as a group OR collection.  So have a button to toggle between the group behavior as it exists now and a collection behavior.  In the collection behavior, I could hit "TAB" or some other button to move to the next control in the collection.  That would make setting the ID's for a bunch of controls a lot quicker.

    I know this is just a silly little feature suggestion, but since this was a recent post, I thought it'd be okay.

  • User profile image
    Maurits

    Sampy wrote:
    A lot of bugs get resolved "No Repro" because we try them on the latest bits and they don't repro.


    Hmph.  Bugs don't fix themselves.  Better to use this strategy straight out of a Microsoft Press book on writing bug-free code:
    1. Find the bits of the version of the software -- as of the reported bug -- in the source history
    2. Build that version
    3. See if the bug repros
    4. If it does, find the bug in that version of the code
    5. See what has changed in that code between the buggy version and the latest version
    It may be that the code changed in such a way that the bug is still there, but the repro steps are different.

    EDIT: If a bug "went away" due to code refactoring -- that is, if the bug is still there but unreproducible -- it can "come back" just as easily.

  • User profile image
    Sampy

    Whenever a bug gets resolved like this, I personally make sure that it's an area that I know that I fixed or the code is gone for.

    I can't speak for all of DevDiv but what you outlined is very similar to my personal process.

  • User profile image
    Maurits

    Sampy wrote:
    I personally make sure that it's an area that I know that I fixed or the code is gone for


    Good for you.  You should get a raise Smiley

    Seriously, diligence in bugfixing should be directly related to pay.

  • User profile image
    brussell

    Maurits wrote:


    Better to use this strategy straight out of a Microsoft Press book on writing bug-free code:



    Ha, that's kind of funny. I just found this Microsoft Press book on our bookshelf at work and decided to read it. It is Writing Solid Code; "Microsoft's Techniques for Developing Bug-Free C Programs". Not sure it is the same book, but sounds like it. I just browsed the Microsoft Press books availble to us online and couldn't easily find a new version. If you know the title of the book you're talking about, I'd like to know if it is newer. This book was written early 90's and talks about 386 and such. Amazing book. Most of the functions and things it details like that are not useful to me, but the ideas are great.

  • User profile image
    Maurits

    brussell wrote:
    Writing Solid Code


    That's it.

  • User profile image
    brussell

    Maurits wrote:
    brussell wrote: Writing Solid Code


    That's it.


    Just thought that was kinda funny for you to mention such an old book that I had just found and started reading.

  • User profile image
    Klaus Enevoldsen

    We are developing a smart client solution in VS2005, we started out in VS2003 and when beta 2 came out we switched. Taking advantage of generics, ClickOnce deployment and TableLayoutPanel (great features, what did we do before?).

    Recently we changed to July CTP and it is much more stable than beta 2, but it is still full of bugs, I've reported a few, but they get closed as Not reproduced. My problem is I don't have time to produce an example to reproduce the bug, and I don't want to share our solution with the whole world Smiley

    The bugs I have reported is reproduced on our 3 machines so it's not just me fooling around, the bugs are there... I feel that Microsoft better find them and correct them - it is their program and their responsibility to deliver bug-free programs. I appreciate that we get to try the betas and CTPs before they are released.

    Previously we have not seen many patches for Visual Studio 2002+2003, I think that this will change for 2005. Too many bugs/suggestions get postponed to Ocas.

    I fell that the framework has been ready for some time and it is very stable. But Visual Studio does some strange things some times...

  • User profile image
    Sabot

    irascian wrote:
    Good to see you back here sabot!

    This has already been discussed (although not very intelligently) in a thread I started yesterday
    http://channel9.msdn.com/ShowPost.aspx?PostID=99525 


    Hi Ian,

    Yep, I'm back!

    I do believe in what C9 is about and I can't let the chimps take over the jungle.

    Back to the matter in hand.

    On the scale of things Visual Studio Beta 2 is just that. A Beta product! It's gonna have bugs in it right?

    Now the point as I see it is, the wider the audience the greater the number of bugs found. In return we get to see the product before it's released.

    Now I think that perhaps some people aren't confident that the bugs will be fixed before the release, then this is another matter. Does providing another Beta going to bring confidence? Yes may be? 

    Personally I believe that managing our expectation by using the feedback system is no different from any other good software company, so should Microsoft be any different? No, but if a reported bug is found in the software after release then just like any good software company I would expect strong questions to asked.




  • User profile image
    nightski

    Klaus,
    I think what they are trying to say is that yes, the bugs were there in Beta 2.  But since then (and before you reported them) Microsoft had already discovered and fixed them.  Hence they were not able to reproduce them on the latest build.

  • User profile image
    bonk

    I kind of trust the guys at redmond that they know what they are doing and are able to decide for themselfes whether there needs to be another beta. Maybe, just need to hope that  marketing aspects do not infulence thatose decisions too much.) I am that condifent because previous versions of that product have been of high quality.

  • User profile image
    Klaus Enevoldsen

    Nightski,

    Microsoft replied that they could not reproduce the bugs in the same build that I am using (July 2005 CTP)

    /Klaus

Conversation locked

This conversation has been locked by the site admins. No new comments can be made.