Thousandapologies

Thank god or gods(never know if you might insult one of them), Tally was there today to help me with some of my longstanding coding troubles. I’ve never done too much server side coding, so most of it is quite new to me; as are some of the tricks that are quite obvious when you know them, but have to be thought of in the first place.

Last week, when I was coding some of the query stuff, I ran into some weird and unexpected problems that could not be debugged on the server. Today, Tally and I decided to set up most of the stuff also locally, clean up the database structure Bashar originally made, adhere to some standard naming standards and furthermore start debugging. Immediately, we ran into some weird problems; to debug we added a main method to the serverclass. It turned out not to be found in anyway by that piece of garbage named “Eclipse”. Even when it was found the stupid program did not run. I was already ready to take my laptop, open the window, throw out the laptop and celebrate a day of non-eclipseness. Bravehearted Tally, however, remained calm under many such problems and told me just to restart Eclipse. Amazingly, that worked *sigh*! Now that we had the server running locally, we could finally run through all the code and found all the glitches and bugs, I, merest of all codemen and lesser geeks, had introduced. Soon we had all of the code running.

Off to another quest through the darkest swamps of php, css and javascript we went. I had found a javascript version of a semacode generator the other day and that finally allowed me to easily generate and afterwards print them. However, to get it integrate with the rest of the server software, a php module was needed that called a webservice, generated an id for the semacode and then use the javascript to generate the semacode itself. That was too much too ask and out of reach for this poor fellow. Also, this task turned out to be a wee bit more difficult for Tally than he first thought. Finally, we got the stupid thing running but still there are some issues. For example, the object the semacode is attached to is described by a name in the database, for instance the not so original “coffee machine” or the most common OTEC form “exploded microwave”. Any way I tried I could not get the stupid name set correctly in the database. Moreover, the code we hacked together suddenly resulted in the semacode to magically dissappear from the print-outs, argg! That’s something for tomorrow.

Tonight, I aimed my arrows at some other part of the code, that now is already finished, but not tested: the metadata service. All kinds of content retrieval and storage should now be easily possible via a common webservice interface. Additionally, I think the code also can be easily adapted for other usages, for example the competence metadata or other parts of the MACE project. But first, let’s see if its working alltogether. Fingers crossed!