TDD

Posted on November 26, 2008

When people develop software they are supposed to write lots of tests to guide the application design, and later use them for regression testing.

However I still don’t see why they are not using an impact analysis tool they will create a test suite with a subset of tests that will verify pieces affected by changes and no other tests. That would save so much time.

In fact I do not know of such tool myself.

SPAM

Posted on November 25, 2008

Some time ago I wrote a rant, about people on 8th floor of Hall building, who take thier shoes off, and put their legs with thier smelly socks on the tables.

Well, by now they are probably graduated, and started a socks advertising compaign. Here is a spam reply I got:

“Thanks for such an important articles about Socks. This is a very needed info about online socks buying.”

It’s pretty good actually. Kinda like google ads. They probably used public API of a  blog search engine to find that post. I think that this is just a start, with more and more sites exposing their content and capabilities via ATOM, REST and web services those spam ad campaigns will be more and more targeted and personal.

The future is already here. It’s just not very evenly distributed.

Posted on November 16, 2008

IDEA

Posted on November 16, 2008

Resharper for Eclipse. I wish ;-/

Teamwork is a heroic actions by one individual.

Posted on November 16, 2008

*Sigh*

Software development mythodology

Posted on November 9, 2008

Internship at IBM

Posted on November 6, 2008

After I droped one course, I had too much time on my hands and started looking for a job. After one week I had three iterviews:  two with IBM and one with Morgan-Stanley.

Morgan-Stanley did not make me an offer yet, but I think that they will.

The good thing about Morgan-Stanley:

• It is a full time job in the downtown Montreal
• I can have a free ride to work
• I will probably get a very decent salary

But I will probably will go for IBM for several reasons:

• I want to have a challenging job
• I want to work for an IT company
• I want to start living on my own.

Also at my 10am interview at Morgan-Stanley, the place was too quiet – I don’t mind working in quiet space, but not for a weeks, months, years!For IBM I had interviews for:

• SWG-0183378 J9 Java Class Libraries – Intern Software Developer position
• SWG-0183444 Rational Requirements Definition Offering – Intern Software Developer position

It’s a 4 month internship with option to stay as an employee or continue the internship for 4 more months (I will try stay for the summer term). Both teams offered me a position. I think I got a pretty sweet deal: I would have a chance to do something outside my conform zone, get paid well and finally live on my own for a while.
Now which IBM offer to accept?
First one is challenging, I will have to work on both java and C code, but this kind of low level development interests me. Plus hardcore developers probably work on VM.
Second one I think is a more traditional Java application development job. But is has an interesting aspect to it – I will have to develop a prototype in Eclipse for my thesis, so it would be a plus if I learn it at IBM, who are the experts on Eclipse. This way I would spend less time developing software and more time on my thesis. Also the tool they are developing if useful for software maintenance a potential topic for my thesis.
I will probably choose the first one, after all eclipse is just another big library, and I can learn it on my own, and the “proof of concept” application for my thesis does not have to be the state of the art piece of software, it just have to confirm my thesis.

Ambient knowledge

Posted on November 4, 2008

Modern version control systems have a concept of “change-set” – an atomic group of files that were created or changed to solve a problem or implement a feature.  If we create a graph and mark the distance between files in change-set to 1, and then analyse all other change-sets and create nodes for files and edges for files in the same change-set, we can then compute the total weight of edges from node to node and perform the grouping of closely related files. To identify potentially relevant but not modified files (change request, specification documents) we can also use heuristics and father augment the graph with set of open files in the IDE and mark them as potentially relevant.

The visual representation of the resulting graph as a picture or a highlighting of related files in the IDE will probably be useful for software developers and maintainers to quickly identify potentially relevant files and subsystems.

Tags

Posted on November 4, 2008

Distributed software development teams often use web forums and mailing lists to communicate, discuss design and resolve problems. That information is often used by maintainers and application users to recover information about the problem. I think it would be a good idea to add a tagging functionality to applications such as mailman to facilitate access to potentially related posts.

© Copyright 0xDEADBEEFCAFE • Powered by Wordpress • Design by Sebastin.