RE: [Axiom-developer] RE: gcl-2.6.8pre on MAC OSX 10.2

[Bill Page]

On November 3, 2006 11:34 AM Gaby wrote:
> ...
> Bill Page wrote:
> | 
> | If you have an issue about the decision to let Tim Daly manage
> | the update of Axiom Silver via his manual updates of his tla
> | axiom--silver--1 archive (with auto sync to svn /silver),
> | then I think that should be treated separately from the issue
> | of the configuration of the source code archives.
> 
> I'm very happy to see Tim is making changes more visible now, 
> than waiting for ages before seeing light. The main reason I
> originally volunteered to maintain Silver is that I do believe
> in "live sources". Tim suggested at the time that he did not have
> time do maintain too many branches. It therefore was a natural
> thing for me to volunteer to take on the job.

Of course your offer was sincere and with good intentions. But since
Tim cant or wont solve his problems with using svn, that would have
meant that he would have had to send his patches to you manually in
addition to recording the changes in his own tla axiom--silver--1
repository. Or alternatively you would have had to learn how to use
arch (tla). Obviously that arrangement was not working.

>  Now, if Tim has more time to do the job, I'm all for it.
> 
> However, I see practical issue here:  
> 
>   (1) first, we should not have one single person as authority to
>       commit changes.  Every contributor we grant write access should
>       commit its own changes to silver/trunk/whatever it is
>       called.  That way, we don't have to wait that only a single
>       person has time and do the job.
>

Yes I agree in principle, however there is a technical problem so
long as Tim insists on using arch (tla): There is no available
automatic synchronization program from svn to tla, but it does work
in the other direction. This would mean that changes committed to
svn would not automatically be made to axiom--silver--1 and would
again rely on Tim to try to manually keep axiom--silver--1 in sync.
This also was not working.

This technical problem could be solved if Tim were to convert to
using another more current repository software, e.g. GIT with which
is can do a two-way sync. (Or of course, if we all agreed to use
exactly one type of repository!)
 
>   (2) second, the review should be public, instead of being done
>       in private.  This helps people to learn possible obscure
>       cases of the system -- instead of the only-one-who-knows has
>       a conversation with himself and commit. The changes should
>       be sent to the list instead of Tim.

Absolutely correct. I am sorry if I was not clear on that.

> 
>   (3) third, I'm having difficulty in following the reasoning that,
>       most people would develop patches against the SVN repo, they
>       will see their patches applied by someone else to a repo under
>       a different SCM, and wait a day and then copied back, to the
>       SVN repo just to see their changes.  That does not strike me
>       as efficient.

In fact I do I agree with you. But projects like Axiom are about
people as well as technology. Sometimes we need to do things to
accommodate other people no matter how awkward that makes the
process.

>       I consider that an extraordinary waste of time and resources.
> 

Well awkward, yes, but I don't see it as a waste of time or of
resources. The most limited resource for Axiom right now is people
with knowledge of how Axiom works. It seems that we might have to
do some extraordinary things in order to retain all of these "eggs"
in one basket... :-)

Regards,
Bill Page.