5 ms·

I sort of understand their reasoning. It's easier to get people interested in a "full" release, than some 3.0.21 version every three mongths. It also was more

by eCa 15y ago

I sort of understand their reasoning. It's easier to get people interested in a "full" release, than some 3.0.21 version every three mongths. It also was more wrong to stay on 3.6.* for over two years. A middle road might be to start a year with a full release and then do point-releases doing that year, so in 2013 we would have, for instance: 13, 13.0.1, 13.1, 13.2, 13.2.1, 13.3.

Karunamon 15y ago
This is a valid, on topic, cogent post. Why is it getting downvoted?
- richbradshaw 15y ago
  Because version numbers are irrelevant, what's important is keeping people on the newest version. "Solving" the problem of large version numbers is like solving the problem 1/0.
  - eCa 15y ago
    At work we have a web application purchased about 18 months ago that officially don't support Firefox above 3.6. There were rendering differences between 3.6 and 4 that were serious enough to enforce that restriction for certain users. Apparently, it has begun working as expected again in FF7 (might have worked in FF6, not sure). The application also uses jQuery 1.2.6 (released May '08) - I guess I'll be the one to QA the migration to the present day. Hence I am not yet a believer in "latest version is the greatest version", in all cases. In principle, yes, but there are exceptions that makes version numbers necessary.
    - Skalman 15y ago
      I'm not entirely sure, but I think Mozilla will continue to support Fx 3.6 until they have a better solution for large companies. I'm not sure what that better solution would be, but I believe they're trying to solve the problem.
    - carussell 15y ago
      Let's be clear. "The version number is too high" is one complaint. "I have compatibility issues across releases" is another. They are not the same. They are not interchangeable. What happened here perfectly exemplifies what happens every time a release is brought up. One of these complaints is brought up, then in the next breath it's switched to the other, as if we're still discussing the same thing. We are not. Conflating the two by bouncing between them multiple times in a discussion, every time the discussion occurs results in massive, unproductive churn.
- starwed 15y ago
  Perhaps because this argument has croppped up, like clockwork, with every release since FF5. It's boring, predictable, and doesn't add anything to the discussion.
ianb 15y ago
Mozilla is dropping the idea that some features are point releases and others are major version numbers. As a web browser there's always only been one web, so if a feature is ready then it's ready and it doesn't matter. Internally there's an actual product that other people might interact with -- and Firefox more than other browsers at that, as the traditional way to extend Firefox is to monkeypatch your changes in. But it was never quite so clear what a point release was there either, and it was getting in the way of advancing the core product. I think simple dated releases would be an improvement (e.g., Firefox 2011.1, 2011.2, etc., or tag them based on month, this one being 2011.10, or 2011v2 or whatever.) But that's basically just a PR move. There's no longer anything really like a point release except in the case of regressions that demand an immediate fix (as happened with Firefox 7).
notatoad 15y ago
the whole point of this scheme is that there shouldn't be point releases. having them throughout the year is just as bad as any other point release. the reason behind going to full numbers is that if you don't have the latest version, you have the wrong version. if there is a 13.1 and a 13.2, it makes it seem like people using 13.1 are still using the latest version. they aren't.
keeperofdakeys 15y ago
Here is a post about where Chrome and Firefox are probably heading http://www.codinghorror.com/blog/2011/05/the-infinite-version.html http://www.codinghorror.com/blog/2011/05/the-infinite-versio.... The biggest problem with updates are regressions. Since Firefox uses four release channels (since the introduction of the Aurora channel), most features get a decent amount of testing before being introduced to stable.

by eCa 15y ago

Karunamon 15y ago

This is a valid, on topic, cogent post. Why is it getting downvoted?

richbradshaw 15y ago
Because version numbers are irrelevant, what's important is keeping people on the newest version. "Solving" the problem of large version numbers is like solving the problem 1/0.
- eCa 15y ago
  At work we have a web application purchased about 18 months ago that officially don't support Firefox above 3.6. There were rendering differences between 3.6 and 4 that were serious enough to enforce that restriction for certain users. Apparently, it has begun working as expected again in FF7 (might have worked in FF6, not sure). The application also uses jQuery 1.2.6 (released May '08) - I guess I'll be the one to QA the migration to the present day. Hence I am not yet a believer in "latest version is the greatest version", in all cases. In principle, yes, but there are exceptions that makes version numbers necessary.
  - Skalman 15y ago
    I'm not entirely sure, but I think Mozilla will continue to support Fx 3.6 until they have a better solution for large companies. I'm not sure what that better solution would be, but I believe they're trying to solve the problem.
  - carussell 15y ago
    Let's be clear. "The version number is too high" is one complaint. "I have compatibility issues across releases" is another. They are not the same. They are not interchangeable. What happened here perfectly exemplifies what happens every time a release is brought up. One of these complaints is brought up, then in the next breath it's switched to the other, as if we're still discussing the same thing. We are not. Conflating the two by bouncing between them multiple times in a discussion, every time the discussion occurs results in massive, unproductive churn.
starwed 15y ago
Perhaps because this argument has croppped up, like clockwork, with every release since FF5. It's boring, predictable, and doesn't add anything to the discussion.

ianb 15y ago

Mozilla is dropping the idea that some features are point releases and others are major version numbers. As a web browser there's always only been one web, so if a feature is ready then it's ready and it doesn't matter. Internally there's an actual product that other people might interact with -- and Firefox more than other browsers at that, as the traditional way to extend Firefox is to monkeypatch your changes in. But it was never quite so clear what a point release was there either, and it was getting in the way of advancing the core product. I think simple dated releases would be an improvement (e.g., Firefox 2011.1, 2011.2, etc., or tag them based on month, this one being 2011.10, or 2011v2 or whatever.) But that's basically just a PR move. There's no longer anything really like a point release except in the case of regressions that demand an immediate fix (as happened with Firefox 7).

notatoad 15y ago

the whole point of this scheme is that there shouldn't be point releases. having them throughout the year is just as bad as any other point release. the reason behind going to full numbers is that if you don't have the latest version, you have the wrong version. if there is a 13.1 and a 13.2, it makes it seem like people using 13.1 are still using the latest version. they aren't.

keeperofdakeys 15y ago

Here is a post about where Chrome and Firefox are probably heading http://www.codinghorror.com/blog/2011/05/the-infinite-version.html http://www.codinghorror.com/blog/2011/05/the-infinite-versio.... The biggest problem with updates are regressions. Since Firefox uses four release channels (since the introduction of the Aurora channel), most features get a decent amount of testing before being introduced to stable.