2005 Web Usability "duh-wards"

Everyone loves the new PostgreSQL website, which we launched about a year ago. Since it came out, we’ve integrated in the old advocacy site, the old developers site, and even converted the email archives over to the new look and feel. I don’t think you’ll find anyone involved with the website who doesn’t think that our web presence is in a better place than what it was at last year. Which is why I loved reading the ”Top Ten Web Design Mistakes of 2005” and seeing how many of the gripes I had long ago given up on about our own site which made the list.

The #1 problem this year was “Legibility Problems”, primarily meaning font sizes that are too small but also poor font colors with regard to the background colors. Given the number of times this one has been brought up on the web list, I still think our font sizes are too small. Some have argued that we use the same font sizes as most corporate websites (like ibm, microsoft, and dell), which the counter argument being that we probably shouldn’t emulate sites designed by corporate web teams specifically engineered to suck all measure of creativity from thier midsts. But maybe that’s not the most convincing argument… on the upside the number of complaints has gone down recently; I don’t know if that’s because the general populous is fine with the fonts or maybe they’ve all just given up complaining about it?

Of course that’s not the only issue we violate. We also violate #2 (Non-Standard Links; differentiating between visited and non-visited links), #5 (Bad Search), and #9 (Frozen Layouts with Fixed Page Widths).

Regarding #5, I actually like our search, but it seems to crash sometimes, and users have complained about it not working as well as pgsql.ru or google, so I guess we violate that one. I wonder if we ought to give users the option to search all three via our search box. If not on the main page (which would be tough to squeeze in) then on the results page maybe.

And thinking about #9, I still don’t see why people are against the idea of at least semi-fluid page layouts. We do this in the postgresql documentation and I think it works great; if only we could get consensus about doing it on the rest of the site.

Happy Festivus

Just wanted to drop a line to say [http://en.wikipedia.org/wiki/Festivus Happy Festivus] to everyone. Luckily my son is only 9, so should make the feats of stregth go smoother.

Upgrading to wikipgedia-0.3

If you’ve installed and started using one of the earlier versions of [http://pgfoundry.org/projects/wikipedia/ wikipgedia], you might be wondering how exactly you’re supposed to upgrade to one of the new releases. This is especially true if you’re stuck on 0.1, since we added the tsearch2 bits in 0.2. Turns out there is no easy way to do it, although the procedure isn’t really that hard; basically you have to dump/reload your database from the old schema into the 0.3 schema. Since this can be kind of a pain, I’ve added a [http://pgfoundry.org/docman/?group_id=1000124 copy of the 0.3 schema file] to the project site that you can use for creating an empty database to get started. At least this worked for the [http://slony-wiki.dbitech.ca/index.php/Main_Page Slony site] that I did the testing on, so hopefully it will work for you.

Working the OSS Conferance Track

One thing we seem to be missing in the OSS world is some type of “conferance meta site” that keeps track of the various OSS conferences, the general topic, where and when they are, and when the call for presentation deadlines are. Right now there are a number of ways you can find out this type of information, but it’s all rather disjointed, especially if you’re into more than one discipline. This tends to be especially bad for database systems like [http://www.postgresql.org PostgreSQL] and [http://www.sqlite.org SQLite], who could be involved in a lot of conferences for other languages like [http://www.ruby-lang.org/ Ruby] or [http://www.php.net PHP], but we’d have to know they are looking for speakers first. Anyone know of a good resource for this kind of information?

John Spencer Has Passed Away

Normally I am pretty anti-hollywood, especially when it comes to the daily travails of any given actor. But sometimes you just get hit with a funk, which is what happened to me after hearing about John Spencer passing away. Certainly I never knew the guy, but I have always been a fan of his work and was really shocked when I got the news last night. Here’s [http://news.google.com/?ncl=http://www.mercurynews.com/mld/mercurynews/news/breaking_news/13426114.htm&hl=en a bevy of articles for those interested].

Shout Out to ChrisKL

Just wanted to give a heads up regarding one of [http://people.planetpostgresql.org/chriskl/ Christopher Kings-Lynne]’s newest ventures, the [http://pgfoundry.org/projects/mysqlcompat MySQL compatability project]. At less than a month old, it’s already garnered its first news headline, via [http://software.newsforge.com/software/05/12/15/1611251.shtml?tid=72 an article on Newsforge]. It discusses Chris’s angle on the purpose of the project, why he got involved, and what it means for application developers. Since I’ve known Chris to be a stud for years having worked with him on the [http://phppgadmin.sourceforge.net/?page=about phpPgAdmin] project, I have to say I am glad to see him doing well.

Time for a Reboot?

From my linux desktop at work:

[rob@camel rob]$ uptime

  2:07pm  up 365 days,  5:28, 34 users,  load average: 42.35, 42.72, 42.85

Wikipgedia-0.0.3 Released (Security Fixes)

We’re happy to announce the release of [http://pgfoundry.org/frs/?group_id=1000124&release_id=377 Wikipgedia-0.0.3]. This release contains a couple of security fixes ported from the upstream project by Sven Klemm, as well as a couple other bugfixes. Normally we probably wouldn’t have made this release so soon, but given the security vulnerablilities, we’ve decided it was prudent to go ahead and do this now. Thanks to those who have [http://pgfoundry.org/tracker/?group_id=1000124 tested and submitted feedback] so far.

OSS DB Rodeo Roundup

Was talking with Greg Mullane last night and mentioned an [http://www.theinquirer.net/?article=28201 article in the Inquirer] that I thought provided a pretty good round-up of the current open source database offerings. One of my favorite parts was in the “company profile” section where it listed for PostgreSQL: 
Number of employees: 0 - PostgreSQL is a project, not a company
That’s so hard for most tech writers to figure out that it’s nice to recognize someone who actually get’s it. But the real reason I brought it up, and the reason I’m mentioning it here, is that it also lists that [http://www.ingres.com/ Ingres] has a current install base of 8,000 - 10,000 active users. We couldn’t really decide if this number was high or low (anyone know how this compares with companies like [http://sybase.com/ Sybase] or [http://www.progress.com/index.ssp Progress]?) but it seemed awfully low to me, just considering that there must be at least this many active installations of PostgreSQL, based on the 1 million downloads of 8.0 and a conservative adoption rate of 1% of downloaders. And that to me is significant because Ingres has managed to hang around for 20 years and is even at this point considered a a multi-million dollar company based on [http://www.networkworld.com/news/2005/110705-ca-ingres.html the recent dump that CA did of the product], so you have to figure PostgreSQL really is a major player in this market. Yeah I know, most of you probably already knew that, but it’s always good to have some empirical evidence you can show to others. And by the way, it looks even worse if you compare it to my$ql’s claimed 6 million installations. While I would bet that the real number of my$ql installations is considerably lower, it’s still high enough to show you just how marginal Ingres has become. If you’re running Ingres, it might be time to start making your [http://www.sraapowergres.com/en/whitepapers/informix_whitepaper.pdf migration] plans.

Wikipgedia Project Update

Haven’t had much time to spend on [http://pgfoundry.org/projects/wikipedia/ wikipgedia] the last month, but found some time to catch up on some things tonight. I fixed a bug or two in the watchlist code, but there are [http://pgfoundry.org/tracker/?atid=534&group_id=1000124&func=browse others]. Also noticed we’re not call-by-reference safe, which would be nice to fix but might be quite an undertaking. One thing I noticed was that we’re really drifting from the mainline code. Part of this is that you got me spending about 4 hours a month on this thing vs. the large wikipedia development community (which as an aside is the sourceforge [http://sourceforge.net/potm/potm-2005-10.php project of the month] in October), but another part is that wikipedia is not even close to a stable codebase; there are hacks and half implementations everywhere [[image /xzilla/templates/default/img/emoticons/sad.png alt=”:-(” style=”display: inline; vertical-align: bottom;” class=”emoticon” /]] Oh well what can you do?