1. This site uses cookies. By continuing to use this site, you are agreeing to our use of cookies. Learn More.

Preferred 921 software release type??

Discussion in 'Standard Definition Receiver Support Forum' started by jsanders, Apr 7, 2004.

What is your preference to release schedules?

  1. Less frequent releases, with lots of bug fixes and features added:

    0 vote(s)
    0.0%
  2. More frequent releases, with just one or two fixes, or features added:

    8 vote(s)
    100.0%
Thread Status:
Not open for further replies.
  1. Apr 7, 2004 #1 of 13
    jsanders

    jsanders Hall Of Fame

    1,191
    0
    Jan 21, 2004
    What is your preference? Less frequent firmware updates with lots of bug fixes, or frequent releases with just one or two fixes? :confused:

    I realize that doing quicker release schedules require more frequent testing (same stability testing load, although less features to test), and such things may not be as easy to do. I'm just curious as to what people like....
     
  2. Apr 7, 2004 #2 of 13
    kmcnamara

    kmcnamara Icon

    860
    0
    Jan 30, 2004
    I work in software development. With more frequent releases, you're just asking for trouble. You tend to rush too much and introduce new problems.
     
  3. Apr 7, 2004 #3 of 13
    jsanders

    jsanders Hall Of Fame

    1,191
    0
    Jan 21, 2004
    The more changes made increases the possibility of unknown problems. Less changes made means what was changed stays more the same. Less probability of more introduced problems.

    In the end result is the same, the question is about how many points are seen on the journey.
     
  4. Apr 7, 2004 #4 of 13
    Mark Lamutt

    Mark Lamutt Your Neighborhood Liasion

    12,527
    1
    Mar 23, 2002
    My feeling is that once the 921 is close to completely stable, with only the occasional hiccup, we'll start seeing releases on a longer schedule - probably once every 2-3 months, rather than once a month. More time will be spent in beta testing and in feature development once we get to that point. Right now, everything that you guys get is tested and speaking from personal experience, many many hours go into the testing, but those many many hours end up being over the course of a week to 10 days, rather than over the course of a month.
     
  5. Apr 8, 2004 #5 of 13
    jsanders

    jsanders Hall Of Fame

    1,191
    0
    Jan 21, 2004

    Now that makes sense! Let's say that the theoretical ideal was reached - no more bugs! We wouldn't expect any bug fix releases. Feature enhancements might come around, but would be far and few between. When the latest version of microsoft office comes out, the average user may find that the new features don't justify the upgrade because the product is mature. When a product matures to that stage, releases don't happen too often.
     
  6. Apr 8, 2004 #6 of 13
    Kagato

    Kagato Godfather

    414
    1
    Jul 1, 2002
    I think one of the things that has really stuck me about the 921 forum is the number of bugs that are present in the 721. It's just the 721 didn't have nearly as much attention as the 921. That's not to say the 921 doesn't have it's own unique bugs and issues.

    The more frequent releases you have the more your costs go up. I would go with a hybrid system. Lump bug fixes into frequent builds. With features introduced in longer builds that have greater time for QC.
     
  7. Apr 8, 2004 #7 of 13
    jsanders

    jsanders Hall Of Fame

    1,191
    0
    Jan 21, 2004
    It would be nice to see some of the "low lying fruit" that is easy to pick, ie., the bugs that are easy to fix to be taken care of in a couple of quick releases. Some examples that would make things a lot easier for many would be fixing the user interface fiasco of trying to set a manual timer. Probably not to hard to fix the one that requires you to set the timer, then edit it two more times to set the padding to zero. In addition the bug that limits you to set the manual timer on certain days, otherwise, the day gets reset during the second edit when you get rid of the padding. Those two can even be given to the 721 code base if my understanding is correct. Maybe adding default 30 minute blocks to the OTA guide to select manual timers would also be an easy thing to add.

    This by no means diminishes the big issues, i.e., the OTA lockup issues. They take longer to solve, and we have to patiently wait for them either way.

    It may be easy to fire off a couple releases to do some of the easy bug fixes. It may be at the sacrifice of more time spent testing rather than coding, and more money spent testing. The added testing may wring out a more solid product though. Hard to say.
     
  8. Apr 8, 2004 #8 of 13
    jsanders

    jsanders Hall Of Fame

    1,191
    0
    Jan 21, 2004
    Has anyone heard of queueing theory? Here is an example. If you are running a network printer, and you have 10 jobs queued, which do you do first? First, examine the jobs. If three of them are just one page letters, and one of them is a 200 page book, you do the 3 one page letters first because you have accomplished 30 percent of your tasks with minimal effort. The 200 page request has to wait for a while for the job to be done anyway, and waiting an extra 3 pages doesn't have much impact. However, the you don't want the 3 people waiting for one page documents to be waiting around while the other person waits for his book to be printed.
     
  9. Apr 8, 2004 #9 of 13
    kmcnamara

    kmcnamara Icon

    860
    0
    Jan 30, 2004
    I thnk your queue theory is too simplistic. The 3 pages are a known quantity and you know how long it's going to take to print.

    Software development would be more like the 1st page came out garbled so you had to go back and figure out whether it was a software or hardware problem which might take a few days. Then test it again locally to make sure everything is still OK. Then try to print again. Then lets say the 1st page comes out OK and then 2nd page (now a week later) prints but gets jammed in the printer. At some point, it makes as much sense to just roll the 3 pages into the 200 and do it together. It's all about risk. Maybe the 3 pages (builds) would come out with no problems in which case the risk was worth it.

    I understand your low-hanging-fruit thought and if these particular fruit are very isolated AND the build process is ironclad, I'd go for it. I don't know enough about how Eldon does things to say.
     
  10. Mark Lamutt

    Mark Lamutt Your Neighborhood Liasion

    12,527
    1
    Mar 23, 2002
    The 921 forum isn't really the place for this discussion, so I'm moving the poll and thread to the Dish DVR forum. I'll edit the title so that it reflects that this thread is about 921 software releases.
     
  11. jsanders

    jsanders Hall Of Fame

    1,191
    0
    Jan 21, 2004
    You're right! Software does take more effort than to print out a page. Sometimes software does garble things up too. The problem I see with your argument is that you are suggesting that pages 1, 2, and 3 will be garbled, and the 200 pages are going to come out just fine!

    In reality, though. There are probably going to be a lot more garbled pages in the 200 page document than with the three one page documents. And, as such, these will take much longer to fix.
     
  12. Stosh

    Stosh Godfather

    301
    0
    Dec 16, 2003
    For me the perfect release would be for a truck to pull into my driveway, the driver brings my new 921 into my house, hooks it up, and leaves without taking my credit card information.

    Oh wait..that wasn't the question, was it? Never mind! :p
     
  13. jsanders

    jsanders Hall Of Fame

    1,191
    0
    Jan 21, 2004

    I'm sure we would all prefer that one! :D
     
Thread Status:
Not open for further replies.

Share This Page