[Linux-aus] The case for an established product to replace MemberDB

[Anthony Towns]

On Thu, Jan 22, 2015 at 03:24:09PM +1100, Kathy Reid wrote:
> There's been some excellent debate here about the future of MemberDB, and
> whether a mature product - possibly CiviCRM, but not necessarily - should be
> favoured over a collection of lower level utilities.
> I'd like to make the case for an established product.

A lot of those arguments seem, to me, pretty much the same as the
arguments people were making for proprietary software ten or fifteen
years ago [0]. I think time's proven they were wrong then, so personally
I don't get them much credit now.

That said, if you were to take the bullet point advantages:

> Maint[ain]ability and supportability
> Alignment with values
> Feature set and applicability
> Cost of ownership
> Legal compliance

then I think that would be a totally reasonable starting point for
criteria to evaluate concrete proposals for replacements for memberdb.

If you want to use it as a way to rule out a bespoke open source
implementation up front, I'm not a fan.

> Maintability and supportability
> While less mature tools and utilities may do the job - and do the job well -
> and even do *exactly* the job we want them to do, they fall down when it
> comes to maintenance and supportability.

You could equally say complicated tools may do the job, but they fall
down when it comes to maintenance and supportability. To take civicrm
as an example: it's not packaged for Debian or Fedora, the recommended
installation method seems to be via unauthenticated http to sourceforge.
Both drupal and civicrm also have what looks to be a fairly regular rate
of security vulnerabilities.

> We have a small team of people who
> work on ZooKeepr - which is an integral part of running linux.conf.au - and
> both getting new people up to speed and willing to commit unpaid time to
> maintenance is difficult. We all have day jobs/lives/other commitments.

Getting people willing to commit unpain time and maintenance is
difficult in any event; but this is a volunteer organisation, so that's
something that has to be dealt with. The reason it works at all, IMHO,
is that there are plenty of people who believe in the principles of free
software and the open source definition, and are willing to put time
into creating and working on software in that model.

> Alignment with values

I think our values certainly include such things as:

 - creating new software to meet our own needs (ie, hacking)
 - making lots of small, reusable tools that work together (ie, the Unix
   philosophy)
 - having access to the source code to software we rely on (ie, freedom 1)
 - avoiding sharing our members' information to third parties (ie,
   privacy)

If you're coming up with policy that makes those things seem like bad
things, I think something's very wrong.

> I feel that it's actually a closer alignment to our values to
> support and adopt an existing mature product - even if we need to pay to do
> so - than to build another tool which will likely entropy over time - as
> MemberDB has.

(I think you mean "atrophy", not "entropy")

MemberDB's over ten years old now; that's actually pretty good value for
the couple of thousand (?) that was invested in it, even at retail
software rates. ($15/mo for civicrm hosting would be about $1800 over
ten years, too)

I'm not sure what options we would have had if we hadn't written our own
software back then; it might have just meant maintaining the membership
list by hand? Certainly that's how elections were conducted up until
that point (ie, with bits of paper at the AGM).

In any event, using projects with an open source community behind them is
definitely a win: it means both that we can make our own modifications
as we see fit (open source), and that we benefit from other users'
modifications each time we upgrade (community).

Comparing that to Zookeepr, I think we've got the following choices:

 - keep using zookeepr, despite no one else using it much

 - switch to another open source conference management system like summit
   [1], pentabarf [2], etc.

 - switch to a hosted conference system like regonline [3]

I think the third option would probably be easier and nicer, but it also
goes directly against the reasons that I support LA. Swapping zookeepr
for summit or pentabarf would be possible, but I think you'd lose about
as much as you'd gain.

The point of free software is that you can do things yourself -- you
take the work that other people have done, and you put it together in a
way that meets /your/ needs, whatever they happen to be. The peak body
representing this idea in Australia really should be living that ideal.

Cheers,
aj

[0] Open source is unmaintainable and you won't get any support. It
    doesn't have the features that you need. Sure, it may cost less
    initially, but you'll pay for it in the long run. Oh, and you probably
    won't be able to meet your legal requirements if you use it.

[1] https://launchpad.net/summit - used by Canonical, Linaro, DebConf

[2] http://pentabarf.org/Main_Page - (formerly?) used by CCC, FOSDEM,
    DebConf

[3] http://regonline.com/ - used by the Linux Foundation