I had a pre-sales consultant at my office today to demo one of the major OLAP players on the market. The graphic capabilities were of course impressive and the functionality to read standardized datasources worked well enough. Looking at the easy stuff is however boring and that the software could do OLAP 101 wasn’t really impressive, I always fail to find questions to which the answers should be given interesting. To see it perform in edgecases I asked him to parse data from my Visio documents, something that I actually need and know is non-trivial. This is where the Microsoft Way started to show its face. Only when saved as XML and with the files renamed to the file suffix .xml did we get somewhere (when will be see Windows software that actually know how to validate fileformats?) but without luck, the correct strings we’re nowhere to be found. How interesting that I did the exact same thing, but made it work, in about 20 lines of Perl the other day..
Next up was digging deeper into the stringhandling of the embedded scripting language. The reason for wanting an OLAP package in the first place is to create logical relations between data of different types in an abstractionlayer where datatype and source is uninteresting, naturally I assumed there would be loads of nice tools to munge data.. and yet again I was utterly dissappointed. The only stringmatching functions available were the usual Visual Basic suspects such as Left(), Right(), SubStr() and friends. When I asked about regular expressions the facial expression was enough as an answer, sigh. When is the Windows community going to realize that regular expressions is The Right Way and their utterly inadecvate crap is Clearly Wrong? Not only is it not portable, it produce hundreds and hundreds of lines of code in order to cover all edge-cases which is a nightmare to write and review.
..but the GUI was pretty and could display all the data I already knew about.. in many pretty colors. wow.