Since I sometimes get bored because I should write my thesis, I proofread other peoples assignments. I know my English is far from being perfect and will probably never be, I still can figure out logic errors. One of them is

carefully estimated exact time required

I can just estimate that this assignment will be handed exactly 5 minutes before deadline ;-) Even if it should have been originally finished on sunday at 3pm

from time to time i just have to move back to this lovely small village. due to the lack of attractions, it is good for getting things done. so maybe the perfect place for finishing my thesis in quite a short time. now i just have to figure out what topic i should choose. ubiquitous computing is just to broad. so i will just think another day about it.

Life has been pretty normal the last days. Thats why there where no really updates on this site.

I have been working on my thesis, playing around with some technologies. I did a presentation about MOF-based Modeling of Abstract Syntaxtrees(in german). It is part of the stuff I have to do for my thesis. The model I created for GastMM was created by hand, but this is not very nice, because now we have to nearly identical models. So I now try to write a tool that make the original model MOF-compliant.

This transformation can be done in several ways:

• Use XSLT. Since I have no idea of xslt, this is not an option
• Use DOM. I used JDOM for it. Its nearly finished by now and looks like its works somehow.
• Use XSD Schema Infoset, a metamodel for XML schema based on EMF. That is what I will try next

I also started playing around with GMF. As soon as I have some news about that I will write here.

After going an a party on thursday afternoon, coming home late i spent most of my weekend working on my thesis.

I now finished most of the Gast metamodel. I should now work more or less. The only thing its missing is

a hole bunch of documentation. But that will come soon.

Sine yesterday there where only fun activities on the schedule, the newbies from my group decided not come. So I had a free day, which I use work on the my thesis. I think that I will just design the metamodel after mof and then use xsl to transfrom it accordingly because the file format it not xmi.

After bbqing I went to tirili. Pics can be found in my gallery

After having the problem of ordered assoziations, 3 different options have be discussed:

1. Define a constrain type like before between 2 assoziations,

2. Define a meta-association from which all the other assoziations go off and the metaassoziation orders them

in the right way

3. Just use tags.

Well, none of the 3 solutions solve the problem because every information added would just appear in the xmi file but does not provide any order. So the only solution is to postprocess that information. So every association will be tagged, because the tag mechanism is allready provided by the mof model, while the other ones would require to define those elements first.

I still have no clue how to sort them later, but I will figure that out.

After trying to provide a MOF-compatible meta model for gast, i ran into one major problem:

how to order association. I don't know why the hell i came up with that idea that will just cost me a lot of time. The main usage is that i can use qvt, but i think i will hack something together because otherwise i will run out of time.

Now i have to face this problem. Using stereotypes is not an option, the are not available in mof, so no use. I could define my

one ones. Since they are defined in the infrastructure of the uml, it should be possible. But actually i don't want to do that.

One idea today was to see if its possible to tag the association itself, but i have to see if mof allows that.

but that will wait until saturday i think. the next 3 days i will be tutor for the new students.