Re: Future of LedgerSMB: Ideas and RFC

[John Locke <..hidden..>]

On 05/18/2011 02:36 PM, Erik Huelsmann wrote:
> To me, this means being able to develop using the customers model,
> differentiating between companies and people. To John Locke this seems
> to mean the a stable Reconciliation interface. What does this mean to
> others?
>
Huh, that's kind of funny, reconciliation is only important to me
because the reconciliation in 1.2 so totally screwed up our books we've
basically had to start over. The reconciliation in 1.2 is the same as
SQL-Ledger -- which leads to very easy clearing of the wrong
transactions, and lots of other issues that have my bookkeepers in fits.

Based on the promise of improved reconciliation, we decided to jump in
and try 1.3. And while we've hit bugs there and in the new voucher
payment/receipt system, overall they are a big improvement. And I do
have more confidence in our numbers, with the extra consistency checking
going on.

But... nearly everything else has been a drastic step backwards in
usability/consistency.
> Of all the people reading this list: what does "getting 1.3" mean to
> you? Do you feel like it's an improvement? If so, why, or if not, why
> not?
Basically, it means getting the current good stuff from 1.3 combined
with interfaces that actually work, at least to the level that 1.2 did.

That means for all the variety of reports out there, that you can pick a
period from the drop-down date selectors and actually get transactions
from that period. That means showing an aging report with subtotals for
30, 60, 90 days, not just one big total for all outstanding. That means
a trial balance or income statement that lets you compare two dates.
That means being able to actually find a customer when you type their
name in the customer search, or have a working sales tax report. When
you hit the email button before posting an invoice, you can fill out and
email and click send, and then get an error and lose your message. After
posting/re-opening an invoice, when you enter an email address into the
field different from what's in the customer record, it ignores what you
type and sends to the customer record.

The hard stuff is done, and I do have more confidence in the underlying
system than I did with 1.2. But for regular users, it's a huge step
backwards -- on pretty much every screen, there are fields and buttons
that do not do what they say they do, do not work the way they did in
1.2, and in many cases do not work at all.

These are all really pretty easy bugs to track down and squash -- but
there are a ton of them, and no issue queue that anybody pays attention
to to keep track of which ones have been done and what's remaining.
These would be great for developers wanting to get involved, learn the
code base, and lead their way up to more substantial contribution,
leaving Chris's time for the good stuff -- but there's no effort to
enlist developers to do this, or to facilitate those of us motivated to
get it working to actually get code into the code base. I've got a dozen
commits in my git repo that haven't made it into the release. I'm sure
Chris is as tired of getting bug reports from me as I am of sending them
to him in lieu of an issue tracker and broader active community.

I feel like I'm alone in using LSMB 1.3 in production -- I keep hitting
really basic bugs that prevent me from being able to get my books done,
and I'm sure would affect anybody else using the system. Again, we're
limping along, patching as we need to and have a moment, getting lots of
help from Chris -- but it's far from ready for production. To get there,
it needs more daring shops who can live with some pain to actually put
it into production, like me, so we can all start squashing these bugs --
and we need a way to get the fixes quickly shared with everyone else,
without Chris being a bottleneck.

David Mora wrote:
> After a year+ of development, and the investment of some resources, I
> decided I needed to drop it. One of the main reasons I quit was that
> the milestones and deadlines  started to get farther. And as someone
> on the list already said, i started to see post about 2.0 while 1.3
> was not even out. So it seemed that all the hard work was in vain.

This captures the problem in a nutshell -- too much bikeshedding, not
enough actually getting working code. I've seen David and at least a
couple other contributors come in, make really substantial
contributions, and then get disillusioned by the lack of getting
something out the door and move on.

David, I'm glad to hear you're still willing to contribute. I hope this
discussion leads to enough change in the project to get a more robust
team put together. I'm certainly willing to contribute however I can,
and I can start by posting the patches we've been making to make it so
we can even use 1.3 to the devel list. I do think github is a better
platform than source forge, because then somebody more knowledgeable can
be an effective release manager and have a lot easier time accepting
contributions from a wide range of contributors while still being able
to easily review the incoming code. But if that's too much of a switch
right now, at least we should all start using the issue tracker we
already have on SourceForge, if nothing else. 4 bugs in there from 2010
(one from me, with no responses). None from 2011. And however the
project wants to do it, find some way to get more contributions going --
I think there are quite a few people on this list willing to contribute,
who are frustrated that it's not apparent how.

And yes, it's insanely hard to just get up and running... let alone
trying to upgrade from 1.2. I think those of us who have managed to get
there should be tackling the interface bugs that make it a usability
challenge, while Chris spends some time making that transition a bit
easier, get more developers to a point where they can contribute.

Cheers,
John