Linux User (& Developer) Expo 2004 - A brief account

This is a brief summary for people - Oliver is busy the next day/two and I'll
be out and about Friday, so I thought it might be sensible to get some notes
down. I'll post another message early tomorrow morning with requests for
action - if anyone can run with these, that would be much appreciated - I've
got 2 days of work to catch up on.

Held in London's Olympia centre, this was Tue/Wed this week. Attended by
Oliver Elphick and Richard Huxton (me). We had about 2m of frontage on the
outside of a small square called the ".Org village" along with KDE, BSD,
Linux user groups, Scribus DTP package etc. Our place was sponsored by
uklinux.net - a linux-friendly ISP.

Oliver had brought a laptop, tower PC with (large, good) flat screen and laser
printer. We had one socket between about 4 groups, so first thing Tuesday
morning the power failed and an electrician had to be fetched. We had no
poster with logo nor any brochures. The laser printer was probably what
tripped the power first time, so we couldn't print any (I had put together a
1-page summary of useful resources). After lunch the show organisers were
kind enough to let me use one of their printers, so I ran off a dozen
handouts (it was an inkjet, so this was a slow process). On the Wednesday,
Oliver printed off a run of summaries he'd put together before the show
started, so we were OK then. On both days, Oliver had a slideshow running on
the flat-screen monitor.

Questions asked (in no particular order):

Q. So how do pronouce [points]?
If anyone has a time machine handy, perhaps we could change PostgreSQL to
something people could say. On the other hand, they stepped forward to ask,
so maybe it's a good ice-breaker.

Q. Why are you better than mysql then? (Not as often as I'd feared)
Sometimes they obviously knew the answer, sometimes were genuinely curious.
Some observed that MySQL AB were much better known, we responded with the
usual about marketing budgets. People seemed quite interested that Oliver and
myself had just volunteered ourselves for this (of course, we both had our
business cards with us too, so it's not completely altruistic).

Q. Who funds it? Who writes it? (a couple of times each)
A few people puzzled about how PG could work. Especially as we made
comparisons to Oracle, rather than other open-source projects. I think this
is partly because a database is a big expensive thing, rather than the sort
of project a teenager might start from his bedroom.

Q. How do you compare to database-X?
Mostly Oracle people showing an interest, then MySQL questions then MS-SQL and
finally one Progress user. The more "business-y" types tended to be looking
to either replace Oracle or were expanding database-usage and wanted to avoid
any further licence costs.

Q. Can we do Y like we do with X?
People interested in porting or looking to start a new project. Two main areas
seemed to stand out - XML and Java. Unfortunately, these two areas are ones
where Oliver and I weren't experts.

Q. Copyright issues with Oracle pl/sql and plpgsql?
>From an ex Oracle employee??? Couldn't see there were any as they're not
really similar enough.

Q. How stable is it? How far can it scale?
Scalability didn't seem that big an issue for the people we talked to.
Quad-CPU Intel seemed to be what they were thinking of as large. Same with
database size - they wanted to make sure we could handle GB rather than TB.
As far as stability goes, the most serious bug I could remember from the last
year/two was that one where you could lock on startup at a block-boundary in
WAL (or whatever the details were - 7.3.2 iirc). Some people looked sceptical
when told that as users we didn't really think about PG's reliability any
more. Suggested they check the mailing lists/release notes and also mentioned
the tender document for Afilias' .org bid (which is hopefully still online,
since I said it was).

Q. Limitations/New features coming up?
I said for me nested transactions was the most eagerly awaited. I was more
likely to forget something on our feature-set than find a problem.
Having a simple to setup replication system might be a useful one to mention.

Pros (no order):

1. A good feeling from those who stopped at the stand.
Those who knew about PG seemed well-disposed to it, those that didn't seemed
quite impressed by what they saw.
Several of the people were clearly senior decision makers in a variety of
companies, and one was from a household name in the UK.

2. PgAdmin
Easier to show people things with a GUI. Some were quite surprised by the
sophistication of it.

3. Oliver's slideshow
Less for any information people got from it, but it gave them something to
look at briefly before making eye contact, or something to read if both of us
were already talking.

4. We were busy
Sometimes both dealing with different people, neither of us had time to wander
around the show. I know a couple of people wandered away because we were
already dealing with other people. I didn't keep an accurate count, but I'd
guess 20 people each day, and some of them took a while to discuss issues in
detail.

Cons:

1. The stand looked empty and we didn't have any brochures.
We were among the worst presented stands on the first day (and I'm talking the
.org ones, not the paid for). Second day our poster arrived and we didn't
look so bad. People seemed not care, possibly because we were both
volunteers, but then if people were put off, we probably wouldn't have seen
them.

2. We didn't know the answers to some questions.
Particularly XML/Java, but I'm assuming the situation would be the same for
any set of people/set of features.

3. Not enough networking
Should have gone round making introductions with all the other stalls. Hoped
to have time to talk to Novell/SuSe and see what their line on databases was,
but no time.

Recommendations:

1. PDFs of PostgreSQL+Logo in A4/A3/A0 (and whatever the US equivalents are)
2. At a busy show, you ideally want 3 people, one of whom can spend time
talking to other exhibitors.
3. A single-page handout: A4, single and double-sided options for cheap
printing so every passer-by can grab one.
4. A short slide-show in OOo format - say 6 sections, 12 pages
5. Six handouts covering each section in more detail - for reproduction in
better quality and to be handed out to those who take an interest in that
section.
6. A volunteer team of topic experts: internals/java/xml/php+perl/...
whatever. Doesn't need to be publicised widely, but I don't even know who I
should ask about our XML capabilities.
7. Funding would be nice, but organisation is more important. We need news
items before and after the show. We need to know what shows are happening. We
need to make sure all printing is ready, and those involved have brushed up
on areas they don't know enough about.
8. We need an elephant. If IBM can have a helicopter at the show I want an
elephant. Oh, and a shovel...

And Finally:

All our plaudits should go to Oliver who lugged a PC,screen,printer and laptop
from the Isle of Wight to London. He also provided the slide-show and printed
off the handouts for the second day. Without him, we wouldn't have had a
presence there, and if we had it would have been me clutching a dozen
handouts, looking forlorn.

Really Finally:

In case anyone is in doubt, I felt the show was a success for PostgreSQL.
There's probably two shows a year in London, and perhaps one in
Birmingham/Manchester. We should probably try to be at all 3 if possible.

--
Richard Huxton
Archonet Ltd