When ndw started at MarkLogic I unwittingly formed a good opinion of the company for no real reason other than Norm being there. I followed Dave Kellogs blog for a while too. More recently I read this piece, a little less enthusiastic about the product. The latter made me realise there was an eval version, which I installed yesterday. No problem. Registered with MarkLogic, got back a sign-on code and started it up.

I then looked for documentation. Mmmm

1. It's mainly in PDF
2. It's pretty well scattered
3. They've used jargon where plain English would suffice, without any glossary
4. It's dated
5. It has (for me) serious omissions

Installation was good. The gs.pdf is the 'Getting Started' guide. My next port of call. This ran through setting up a new server, loading minimal data and querying that data. Great. Except the xquery code given doesn't work 'out of the box'. Seems you need to know all about the various versions of xQuery supported by MarkLogic. 1.0 strict, 1.0 standard - with extensions (think about that one!) and The May 03 draft. No mention of this in the 'getting started guide'.

A little poking about in The rec and I resolved the bugs, thinking to myself why? Isn't this the stuff that new customers are supposed to read to get to know the product? As well as the curious who may buy? Not much of a recommendation for document quality.

Having run the demo I started to think about loading my FAQ into the database. A few hundred XML files, each having a few tens of docbook qna entries. Except xquery doesn't support reading a directory full of xml files. In ignorance I wrote a piece in java to read the files, hoping to call that from xquery file. Oops. MarkLogic isn't written in java. Back to Google.

This seemed the best bet! (Big mistake). Eventually loaded the jar files, added them to the classpath and created a script. Missing class. Seems MarkLogic like to confuse? They have (at least) two versions of xcc.jar on the site. The one pointed to from the howto / tutorial isn't right. But then again the other one, here... or more precisely, go to the bottom of the page and you're redirected to here, then here.... doesn't work either.

At this point I gave up. A few weeks ago I tried Orbeon, which loaded just as easily... let me load all my xml files in about ten seconds and was about as simple as it gets.

Sorry Mr Kellog, your documentation sucks so I'm in no position to judge your product. I'll let you know if it improves.