Re: Future of LedgerSMB: Ideas and RFC

[Luke <..hidden..>]

On Tue, 17 May 2011, Chris Travers wrote:

> For the last few years, LedgerSMB has achieved significant growth.
> Some of that growth has come at an organizational cost and for that I
> apologize to the community.  Now I have to try to help put the
> organizational stuff back together.

We had the start of a discussion about that on the devel list a few months 
ago/near the end of last year, did we not?  There were some good ideas 
there, although my own thinking may have revised a little since then.

> I think the major priorities at this point need to be:
> 1)  Getting 1.3 out the door.

Agreed.

> 2)  Focusing heavily on community building

Certainly.  There's not much that can be done for the user community until 
after 1.3 is fully released, and maybe not even until 2.0, but for the 
developer part of the community, see John's earlier message.

 > 3)  Trying to build partnerships with other open source business
> projects (perhaps GNU Med and others?)

Again, I agree, but what do we have to offer them?  Your step 1, of beta 
releases, doesn't solve the momentum problem.  It may have, at one point, 
but right now they would be partnering with a promise.  I wouldn't, were I 
them, unless they are in the same boat.
Right now, our only offering, is a retooled version of 
SQL-Ledger, and a sort of working beta of our own version of the same 
thing, which we say is better, but very possibly isn't.

> To this end I would like to tentatively suggest the following:
> The first is a regular beta release schedule for 1.3...  Maybe every
> other Tuesday?

Sounds good.  It would sound better, if it was every two weeks, which 
would reset the clock every time you released one, thus allowing them to 
be released more frequently.
No point in releasing a beta, if you fix it three days later, and still 
have people downloading and testing the old one, just so you can stick to 
a good, but arbitrary, schedule.

 > There are some committed fixes for 1.2 which have not made it into a
> release.  I would like to release this as soon as possible.  However,
> given the fact that bug reports have slowed, I think it is likely that
> it is not likely that 1.2 will see another release absent developing
> problems  like issues caused by new versions of Perl.

Until you release a full version of 1.3, it is likely that you'll still 
have many users on 1.2.  I'm sorry to say it, but because of the track 
record, holding out for 1.3 and not bothering to fix bugs found in 1.2 
after the next release, seems like just another way the user base will be 
disinclined to support the project in the future.

I'll say it again.  Until 1.2 can actually be replaced with something 
full, that works, that isn't an iffy prospect, and that isn't a beta, the 
only real way to keep users (not developers, or people like me with a foot 
in both camps) interested, is to maintain the flagship product (I.E. 1.2) 
up until the last minute, as if it was the only product.  That's annoying, 
yes, but I think necessary.

You think 1.3 might get out of beta soon, then great - none of the above 
matters.
But I believe you have thought that before, and life has a way of changing 
what we think into "oh yeah, remember when we thought that?".

Saying that 1.2 will probably not see another release after the next one, 
serves only to make those of us who trust it, and more importantly, don't 
trust 1.3, wonder if maintaining 1.2 as a separate project, might be a 
good idea.

(That was proposed privately to me, last time you said something about 
stopping updates for 1.2, by someone who is more than capable of doing the 
job.
I don't actually want to see that happen, because it only serves to take 
away from the future of the project (which is 2.0).  But the more noises 
that are made about ending 1.2, before 1.3 is out the door, the more 
those who think along those lines, will have cause to do so.)

> But in addition I would like to see what the community thinks.  What
> do you think we need to do to pull things back together and bring the
> project to the next level?

Fix *all* install problems with 1.3.
No other development matters until that happens.  If it breaks again with 
a future beta, then fine, but so far every other person who posts about 
having installed 1.3, has a problem doing it, or immediately after it.
(I am generalizing, of course.)

What's left to do on 1.3?  What things don't work but need to before a 
release?
What things can be fixed after a release?
What features need to be added?

If you have to, relax your standards for 1.3 a little.  If there are 
features which are proving difficult, and they aren't required for major 
operations, disable them in the official version, announce them as pending 
and do subreleases to add them.

(Obviously, I'm not exactly up to date on what is and isn't working in the 
latest versions of 1.3.  Last time I pulled a copy was probably over a 
year ago.)

Luke