Upto March 26th 2008… March 26, 2008
Posted by albaroud in Uncategorized.add a comment
CDLG:
Much more time has been spent on this tools then I would have imagined:
- A large proportion of time went to error checking and tool integrity. The tool itself has crashed quite a lot in the last few months, for the following reasons:
- users have been feeding in a lot of obscure CD requests; mostly with character sets I have never seen before. Various control characters and character sets have an affect on the CD generation, and cause CDLG to break
- Corel Draw opreational erosion: Over time, as CDLG uses Corel Draw, the program becomes corrupted (BadInvoke() error codes). Why this happens, has to do with problems at the OS level which I cannot answer, given my current knowledge of computers thus far.To rectify these problems, I have proposed to ask users to send in there erroneous spreadsheets; I can code in extra protection if I cataloger the characters which cause CDLG to crash. As well, I have developed a simple, ingenious regeneration loop for Corel Draw. When an Invoke() code error is detected, CDLG will destroy CDLG via a .bat file, and just make a new one. This solution is simple and effective in increasing CDLG’s lifespan, for the time being.
- users have been feeding in a lot of obscure CD requests; mostly with character sets I have never seen before. Various control characters and character sets have an affect on the CD generation, and cause CDLG to break
- Corel Draw opreational erosion: Over time, as CDLG uses Corel Draw, the program becomes corrupted (BadInvoke() error codes). Why this happens, has to do with problems at the OS level which I cannot answer, given my current knowledge of computers thus far.To rectify these problems, I have proposed to ask users to send in there erroneous spreadsheets; I can code in extra protection if I cataloger the characters which cause CDLG to crash. As well, I have developed a simple, ingenious regeneration loop for Corel Draw. When an Invoke() code error is detected, CDLG will destroy CDLG via a .bat file, and just make a new one. This solution is simple and effective in increasing CDLG’s lifespan, for the time being.
– the next big issue is Accessibility standards for the tool. CDLG failed, with flying colors. Its not compliant with IBM’s accessibility standards at all. This has involved about 3 days of work, making it compliant…going though page after page, correcting everything in site. This is going to take a lot longer then I thought.
– The code has been cleaned up, and is more efficient then it used to be.
- error prevention, and documentation still need to be done.’
– A new WAS error has been discovered. If the java.exe is shut down improperly, the .pid file in logs is not deleted in time – if this happens, the server fails to restart (and wonderfully, gives NO ERROR MESSAGE TO WHY THIS OCCURS). After a lot of research and messing around, I discovered this. I shall never forget, lest I have to waste many hours again.
SSG:- not much work has been done on SSG so far. Tomorrow, work will start on this tool again.
Eclipse UI Project:
The eclipse project was done 2 weeks ago. Below lists the major events, and what could have been improved:- a major thing noticed early on was, that I wasted too much time trying to manage the project. In the end, it was learned that a CVS repository existed the entire time. Any, more simplistic back up management techniques were in vain. I was resistant to change as well; which caused a fair amount of stress before I accepted the fact and moved on.
- Higher development standards were forced upon me (and I took them unwillingly). I soon realized that, the level of work required was much higher then what I had become accustomed to. Being a student working on internal tools – in a department where everyone is really busy – standards are high, but certainly not perfect. For this project however (one that was presented at EclipseCon 08′), everything had to be essentially perfect. It was implied that any imperfection would be picked up by industry morons/pundits during the EclipseCon presentation.
- An appreciation for User Interface Design was gained. Working with seasoned people in this field, it was a tad mind-blowing how much detail, thought, and preventative programming went into the Interface of this project. As a developer, User Interface people have to be some of the most meticulous (and sometimes, annoying 
) professionals in the field (Testers still give them a run for there money, however).
) professionals in the field (Testers still give them a run for there money, however).- The project ending (like many), held the most intense moments of this entire ordeal. It was 2pm on a Tuesday (deadline), and I was told that a lot of fixes needed done on the errors view of the project. Minus breaks, it took me about 6 hours to get things cleaned up. The code involved an non-friendly mixture of CSS pseudo-classes and JavaScript; I finished about 12am midnight, rather spectacularly.
- Overall, despite some of my negative comments regarding this project, it taught me the most (with respect to the amount of time I spent), compared to any other project I have done in the last 11 months. By far, this project is the closest thing to professional work I have ever done, so far.
Next post: final days, server migrations, end, summer plans.
