Re: [ jEdit-devel ] Structure of the JEdit (OO Design)

[Andy S. <an...@ru...>]

On Wednesday 07 June 2006 09:18 pm, Alan Ezust wrote:
> Speaking as a PhD student who basically dropped out of my program due
> to spending so much time fixing bugs and not enough time doing
> research, it hasn't worked for me yet...

Yeah, I was in a PhD program too -- and within 18 months I knew I could not 
resist  the money to be made with an MS.  That's one reason I appreciate 
those who hang in there:  they sacrifice income (forever) for the chance to 
pursue basic research.  Looking back after 20 years, I think they make a good 
choice!

> I think that it would be really nice if a PhD student could work
> towards a PhD and fix jEdit at the same time, but how can that be
> considered an "original" thread of research? 

People contribute in different ways.  Hayden's contribution is just of a 
different nature.  Great software is not made just with hacking and patches.

> It's a practical problem, 
> but can a solution in jedit's case be considered itself a line of
> research in the area of software engineering, deserving of some
> funding?

There is a way.  You have to hook the interest of a professor who knows how to 
write research grant proposals -- those guys abound, it's their bread and 
butter -- then they can assign graduate assistants to work on bits and pieces 
as part of some larger project that produces the published papers that keep 
them alive.

Ways jEdit might be used in research:

* lots of projects need a plain text editor as a front end.  Lot's of 
academics produce new languages that need an editor.  jEdit's plugin 
architecture is a great solution.  If we make it easier for people to create 
plugins (more docs, clearer examples, wizards), jEdit would be even more 
appealing.

* an example of a nontrivial, mature, widely-used opensource project (which is 
what attracted Hayden)

* a subject for refactoring research (one of Hayden's interests)

* a subject for anyone interested in development environments, especially 
cross-platform ones

* as a community, jEdit is a good subject for people looking into the 
psychology of computer programmers (if anyone remembers Jerry Weinberg's 
seminal book with that title from the 1970's)

If people in jEdit community have ties to research universities, they might 
shop these sorts of ideas around.

What would make Hayden's report about jEdit more powerful to this community 
would be packaging his analysis tool as a jEdit plugin.  (and using 
OpenOffice Calc instead of Excel.  A bit odd working with all that opensource 
software then using an MS product instead of a perfectly fine opensource 
equivalent.)

Andy