Re: On what we want to support: infrastructure?

Josh Berkus wrote:
>
> To refresh, here's the list of everything we might want to spend money on:
>
> 1. PostgreSQL.org infrastructure (servers, bandwidth, sysadmins, SSL, etc.)
> (unlikely to need money, but if it does, the highest priority)

There's one area regarding postgresql.org infrastructure that I think
we could enhance that would benefit the community pretty greatly.

Could the PostgreSQL infrastructure itself serve as an example of
best practices and/or interesting use of PostgreSQL --- and have
the source for the apps poswering postgresql.org be available on
pgfoundry/gborg?

I can think of a couple examples:

* Is search on the postgresql web site or docs or mailing list
archives powered by tsearch2 or some sgml indexing feature
or some other postgresql cool feature I don't know about?
Could it be? Could we see how it's done? I think the docs would
be especially interesting if it indexes the sgml; and I think
the mailing list archives are a pretty nice example of a fairly
large scale search database.

I've been asked why PostgreSQL.org's search apparently uses
ASPSeek and ASPSeek's docs claim the supported database are
"it can be mysql or oracle8 for now." (Though I've been told
in postgresql.org's case it's actually backed by PG, that's
not obvious anywhere.)

* I seem to recall a developer map somewhere. Was it generated
by PostGIS? If not, note that pretty impressive maps can
be generated from PostgreSQL/PostGIS like the links below [1,2,3]
All of these are dynamically generated (change the mapxy or
scale parameters if you don't believe me) from a 90GB postgresql
database of individual road segments; and it works pretty well
and IMHO would make a pretty nice demo and example of how to
use that feature?

* Is the postgresql.org adserver powered by postgresql? That
too would be interesting to many small site webmasters if
the source were available.

* Is postgresql.org itself a database-backed web site? How
about showing the source for that on pgfoundry as an
example web site?

I'm sure there are many other areas that might be interesting; but
does the general idea sound reasonable -- fund using the infrastructure
to show rather than just tell about how postgresql can be used?

[1,2,3] (Some nice maps dynamically generated directly from a postgresql
database that I think could turn into a nice example on the web site)
[1] http://64.127.105.202/maps/mapserv?mapxy=-122.39+37.79&scale=10000&layer=land&layer=roads&map=cp.map&mode=map
[2] http://64.127.105.202/maps/mapserv?mapxy=-122.39+37.79&scale=100000&layer=land&layer=roads&map=cp.map&mode=map
[3] http://64.127.105.202/maps/mapserv?mapxy=-122.39+37.79&scale=1000000&layer=land&layer=roads&map=cp.map&mode=map

>
> 2.a. Speaker travel to key conferences.
> 2.b. Membership and participation in standards and benchmark bodies.
> 2.c. Developer tools (hardware & software)
> 2.d. Performance/testing tools (mostly hardware & hosting)
> 2.e. Development of PostgreSQL code
> 2.f. Porting other OSS applications to PostgreSQL
> 2.g. Printing Marketing collateral for PostgreSQL (CDs, flyers, case studies)
> 2.h. Developing marketing collateral for PostgreSQL (hired writer)
> 2.i. Generally booth duty expenses for conferences (food, signs, internet,
> etc.)
>
> 3. Commercial booths/pavillions at large conferences
> (only if we have money coming out our ears)
>
> I *think* everyone is in agreement on (1) and (3). Where people are arguing
> is for 2.a-i, where people want to set some priorities.
>
> Personally, I don't think that we can set any meaningful priorities for
> categories of expenses in the abstract, which is why I'm pushing a "bang for
> the buck" evaluation. However, a couple of people have pointed out that
> we're still vague on what constitutes "bang". For example, what are our
> comparative criteria for:
>
> a) reaching potential new users at OSS conferences?
> b) reaching potential new users in South America, Africa and Asia?
> c) reaching "suits"?
> d) reaching governments?
> e) developing new PostgreSQL features?
> f) improving standards compliance and certifications?
> g) improving performance?
> h) adding to the number of PostgreSQL OSS user applications?
>
> If we have to compare, for example, sending David Fetter to a Venezualan
> conference sponsored by the government where he will speak to an audience of
> 300 people against offering a prize to Joomla developers who port add-ins to
> PostgreSQL, which "bang" is bigger?
>