📖Glamorous Toolkit Live #4: Moldable analysis of Java systems

the primary idea why glamorous toolkit was started is that developers spend most of their time reading code

reading code is often implicit—nobody talks about it, so it is not optimized

software engineering as decision-making business, not constructing business

(re: code as database)

visualization of class (interface, attributes, links, etc)

• (reminds me of Deep/shallow modules)

constructing visualization of class usage

select sub-expression to examine it

(modern IDEs highlight unused classes and also detect groups of unused classes. this doesn’t cover the whole demonstration but it’s not that bad)

if “reading code” is implicit, you do just that—read. However, if you start looking into it, you’ll notice that there are very different specific activities that you’re doing. → Reading code is not a goal

after you have implemented the analysis, you can create a constraint and execute it automatically (e.g., in CI)

combining constraints into a report with statuses/etc.

gt toolkit seems to mostly use text files and views just enhance them

• (this is similar to how Emacs/org-mode operate, although gt is much more powerful with visuals)