Monday, May 27, 2002

If you want to keep up-to-date on the Mono Project and you use an RSS feed reader, Miguel has added an RSS feed for you. Update: Sam noticed that the feed wouldn't show up in IE. If I had a guess, I'd say that the MIME type isn't something that IE was prepared to deal with properly (should be "text/xml", obviously). I'd further guess that the RSS readers just don't care what the HTTP MIME type is, and assume it'll be validly formatted RSS. [The .NET Guy] 12:41:41 PM      comment []  trackback []

Miguel de Icaza has an RSS feed for updates on the MONO project. To subscribe to it in Radio, click here. (Of course Radio must be running for this to work.)  [Scripting News] 12:41:27 PM      comment []  trackback []

Perl Application Server: http://pas.sourceforge.net/ These guys are writing a proxy server to capture clickstreams for playback during debugging/testing (among other j2ee-type things). I've hacked simple tools like this together before... but it'd be interesting to see what they come up with. Their architecture seems highly Java-inspired, I wonder if they've heard of p5ee? 9:40:52 AM      comment []  trackback []

Mono I sent an email to webmaster@gomono.com to see if they had set up an RSS feed for the site. Miguel de Icaza answered back that while they hadn't, it was onlly because they hadn't looked into it yet. He asked for a pointer on how to set one up, and I sent in some perl code to parse the RSS from the HTML (see http://www.perl.com/pub/a/2001/11/15/creatingrss.html). I don't really know Miguel and it was impressed by how much webmastering time he must spend responding to emails like mine. A couple days later, mono came out with their rss feed, which was cool. Their rss feed ( http://www.go-mono.com/index.rss ) is slightly different from what my code produced, so I think they rewrote it. For Ximian, that would be in character :-) Since I had the code handy, I did the same thing for Lars. 9:18:48 AM      comment []  trackback []

Knowledge Management I had the opportunity this week to sit in on a friend's case study presentation for a knowledge management (KM) product rollout. I have considered myself one of the masses who just don't get the difference between 'true' KM and the search/document mgmt/project mgmt/portal/cms/etc. pretenders. My friend's company does a lot more management consulting than technology consulting, and I think that's the key ingredient missing when I consider these product's differentiation : there really is no huge meaningful difference between these products from an pure technology implementer's perspective. They all store documents, catalog them and provide search functionality, and enable workflow and permissioning to a greater or lessor degree. They all compete with the file system and Outlook (which is my ##@$% KM system). The cross-disciplinary aspect of KM (like OLAP and LDAP below) wasn't something that was ever made obvious to me until listening to the case study. In this, it seemed way more important to get the management on board and make sure that the right data was preloaded, the training and page templates reflected what people actually needed to do with the system, than it was to actually install the software and get the custom code written. As consultants, we've all seen jobs where those items were important, but in KM it almost seemed like that was the only job. The project had on staff a 'change-management consultant': someone specifically tasked with encouraging the adoption of a new way of doing things. I know that this is the kind of icing that clients of modest means are going to balk at paying for, but I think for KM (and probably CMS) rollouts the risk of not addressing the non-technical implementation challenges definitely swings the balance in favor of working with a specialist. 8:54:20 AM      comment []  trackback []

The Why's and When's of OLAP OLAP is almost always about doing financial calculations or projections that are nearly impossible to do in SQL. Typically you throw sales #'s in the database, or maybe clicks/per/page (in a web analytics system). You then break it down by geography, sales-person, branch-office, month, promotion, product, etc. (These are called dimensions). OLAP products pre-aggregate these numbers, so that any time you ask to drill into numbers (month->day, state->city, etc), you're given a subsequent response time. Analogously, RDBMS have indexes to bring back data faster, because you're mostly interested in pulling back one row out of the millions in the database. OLAP 'Cubes' have aggregations (sums, counts), because you may be interested in summing a million rows, but a different million each time. RDBMS OLAP indexes aggregates attributes numbers single rows or search results aggregates for 000's or 000,000s of rows 8:18:11 AM      comment []  trackback []