Ionrock Dot Org

My last semester I got pretty serious about using a versioning system. My software of choice was subversion. I tried arch for a while but I found it extremely difficult to integrate into my workflow. I am sure with enough work I could have gotten used to it, but I was not willing to lose productivity when other tools were available. I am convinced thought that had I stuck with it, I would have become a arch zealot. Its design is really nice and the paradigm it takes seems very logical and extremely helpful. Unfortunately, it wasn't practical for me.

With arch out the window and the masses consistently complaining that CVS is crappy, I started working with subversion. It was still very hard to get in the habit of working with subversion but overall it became rather nice. The biggest boon was that I could check things out on whatever machine I was on, work and check changes in. Of course this is the obvious benefit of a versioning system, but after trying many other systems of keeping a central workspace, this worked very well.

Now that I am working and using CVS every day, I am glad I took the time to learn what I have. I haven't really seen the difference between CVS and subversion but I have "felt" the differences. CVS "feels" kludgy and unstable. I know our sys admin is not a fan so it might be an administration issue, but overall, subversion just seemed easier to use. Of course beggars can't be choosers, so I will continue to get into the CVS groove as much as I can and wait patiently for the sys admin to migrate to subversion (which is in the works I understand).

I usually get along well with others... even CVS.

At my job I work with XML almost all the time. It is a lot of fun because XML is so general that there are endless possibilities for it. When I was looking for a job I considered where I would like to work within the context of open source software. The area that seemed closest to my own interests dealt with notification, which translates more generically to your computer initiating some sort of action. The problem with this is not the medium or the message. Both of these can be tweaked and massaged to fit within anyone's workflow comfortably. The bigger issue is that it is a pain to tweak and massage some program to give you helpful reminders that do not affect your ability to do work.

My solutions ended up looking at the nature of the request and handling the kind of notification accordingly. I invisioned a smart reminder that could be triggered by dragging and dropping a document or email and simply entering a date. I also thought about emailing my reminder service and letting it read the details of the email to create some sort of reminder or notification. The biggest hurdle in this was working with the vast array of documents and technology that one comes into contact with on the desktop. This morning I considered how technologies such as RDF and semantic documents make this much more feasible.

The benefit of using a sematic document format is that new information can be gathered about the document without explicitly stating it. This means that a computer could look at some document you wanted to be reminded of and notice that it stems from project that you have been working on. The project is worth a great deal of money and the deadline is coming up and the computer could see this. The notification then can be sent to your phone and there can be a more disruptive notification on your desktop because all the information pointed to it being helpful.

I am not sure this makes total sense at this point, but I am going to think more about it and see what happens. I am pretty sure that at some point I am going to want to go to grad school and this seems like a pretty nice project for something like that.

In other news... I am going to PyCon!