(no subject)

New Message Reply Date view Thread view Subject view Author view

From: Jonathan Guyer (jguyer@HIS.COM)
Date: 14 Dec 99, 03:33 EST


From: Jonathan Guyer <jguyer@HIS.COM>

At 6:26 AM -0500 12/13/99, Mr. Noyb wrote:

>I have changed a folder name which contains an Alpha file
>(folder and file names are letters & spaces only) which:
:
:
>Is there a work-around for this?

Not until Alpha 8 (and it appears that it hasn't even been fixed there,
yet, although I thought it had been).

No, there is no schedule for release of Alpha 8.


>While I'm asking questions, why is it necessary for so many versions
>(beta & final) of Alpha to be (pre)released? There are some obvious
>disadvantages to this from my POV:

I don't entirely disagree, and I take a rather more conservative approach
to version updates than Vince does, but I want to correct a misconception.
Remove the parentheses in "(pre)released". _Everything_ we're talking about
is prerelease code. In the last six months, _releases_ of Alpha have been:

= 7.2.1 released  last update: 11/10/1999 {5:27:15 pm}
= 7.2 released    last update: 24/5/1999 {6:29:13 pm}

That hardly strikes me as frequent.

If you dislike instability (and who doesn't?) then stay out of the
PreRelease directory.

I see no need to apologize for frequent changes to the PreRelease code,
however. This stuff is inherently "use at your own risk" code and is there
to give bleeding-edge types an early look and us some feedback. It's not
intended for general consumption. The only thing I'd like to see is better
feature stability in the betas and _absolute_ feature stability in the
final candidates (this may involve a more formalized "alpha-Alpha" phase).

>Frequent changes introduce
>instability to Alpha - an intended bug fix has often introduced more
>bugs, for example.

And this is why these go in the PreRelease folder. I'll grant that in years
past, code got finalized before it was anywhere near stable. We've been
much more disciplined about this, lately, though. There's undoubtedly room
for improvement.

>Far more disturbing for me, the behaviour of some
>of Alpha's routines is (seemingly arbitrarily) changed between
>releases. Twiddle is a good example.

Twiddle is a good example of what? On perusing the last year of Alpha-D,
the only discussion of twiddle that I find is a rather lengthy exchange
between Juan Falgueras and Vince about a proposed change of Juan's. Nobody
else said anything one way or the other and Vince implemented the change.
That was in June. David Craig questioned this change in October and,
apparently, nobody answered him.

I won't get into whether the change was a good one or not (I don't twiddle,
so I don't know), but it was hardly arbitrary. Juan explained in detail
what it was he wanted and posted all the code to do it.

__________________________________________________________________
  Jonathan E. Guyer
  <http://www.his.com/jguyer/>


New Message Reply Date view Thread view Subject view Author view

This archive was generated by das@users.sourceforge.net with hypermail 2b29 on 31 Mar 01, 11:51 EST