It is not a beta [LWN.net]

It is not a beta

Posted Mar 7, 2008 16:50 UTC (Fri) by JoeBuck (subscriber, #2330)
Parent article: The first Gnash beta is out

The folks behind this announcement have no idea what "beta" means.

Traditionally, alpha testing is testing by insiders: developers and people internal to the organization that is developing the code, with lots of hand-holding available.

Beta testing is end-user testing, where you're asking non-developers to help test code that you think is nearly ready for final release. No beta announcement should ever contain a line like

"As heavy bug fixing and development is going on, please make sure your bug still exists in cvs HEAD before reporting."

That's because the kind of testers you want for something like Flash have no idea what "cvs HEAD" means. Only developers do, and developer testing is alpha testing. A beta test would require the publication of installable .rpm/.deb packages that non-developers could try out.

I think it would be more accurate to call this the first alpha release. Usable libre Flash is the only think keeping my daughter from being a saint in the church of Emacs. But let's not fool ourselves; there is a very long way to go.

to post comments

Why not?

Posted Mar 7, 2008 19:56 UTC (Fri) by khim (subscriber, #9252) [Link] (2 responses)

No beta announcement should ever contain a line like "As heavy bug fixing and development is going on, please make sure your bug still exists in cvs HEAD before reporting."

Millions of people out there know what the "cvs HEAD" is. Tiny fraction of them are gnash developers. May be one per thousand but probably less. How the heck will you call version which is ready to be tested by someone who is a software engineer but not developer of gnash? This is not alpha - people outside of usual circle (potentially millions) are invited. I'd say it's kind of private beta: version not yet intended for general public, but good enough to be tested by millions. The fact that selection is by knowledge and not by some other random characteristic is irrelevant.

P.S. Not so long ago the fact that anyone who's not familiar with gcc or cvs will even try to use Linux was laughable. Times are changing but IMNSHO it's still not too much to ask from average Linux beta-tester: Jor Averages who don't know about cvs can wait for a later beta or a release...

Why not?

Posted Mar 7, 2008 20:02 UTC (Fri) by JoeBuck (subscriber, #2330) [Link]

I'm all for asking for a larger tester community. I still don't think that they've reached the level of stability that the word "beta" normally indicates.

Cloning Flash is a very tough job, and the people that take it on have my respect. But there's a danger in over-promising; people try it out and then go away.

Why not?

Posted Mar 7, 2008 21:15 UTC (Fri) by jhoger (guest, #33302) [Link]

Two terms I've heard used for that is "engineering beta" or "internal beta." Good enough for
engineers other than the original developers, but not users.

-- John.

It is not a beta

Posted Mar 7, 2008 20:32 UTC (Fri) by debacle (subscriber, #7114) [Link]

Hm, I used to know what CVS HEAD was, but I can't remember. Wasn't it sth. like SVN trunk? Or,
no, well, I'm confused about relatively new projects using CVS.

It is not a beta

Posted Mar 7, 2008 21:36 UTC (Fri) by jwb (guest, #15467) [Link]

Even hard-core hackers will be put off by the requirement to test the CVS head before
reporting a bug.  Gnash developers should treat all bug reports equally unless they know that
a commit later than the last release has likely fixed the problem.