It's a solution in a search of the problem, I think.

If it's a poorly designed system, then the view will be a muddy web of connections. On the other hand, if it's a well designed system, it will be highly modular, with each module being small enough to be readily comprehensible in one go and with very few connections between the modules. So I basically don't see a use for the system that you are describing.

Well, I think the muddy pool of connections is what this is a solution for. While I see it as practically impossible to implement, it would be great for huge legacy systems that grew rather than were designed. For instance, all of a sudden my patient certification table is getting weird records appearing in it. I would be able to find that table in this cloud, filter to connections to objects that have changed recently and individually inspect each one to see if it was the source of the problem.

The big problem is automating this to a degree that it's not a full time job to maintain.

A weighing scale is a solution in search of a problem. Most people will do nothing, and the few that truly care will do so whether they have one or not. So I basically don't see a use for it.

---

If you believe that we improve what we measure then this can be useful.

I deleted a line from my original post before submitting it which said it was a problem I wanted a solution for so that might make it worth persuing!

I like your suggestion that it should be highly modular, not sure what makes that mean its not useful though?

I disagree.

If the objects are generalized enough, this could be applicable to any industry with complex, integrated procedures.

Rational (IBM) have tools that can do it for C/C++ - what language are you looking for ?

What you're describing sounds a little bit like LightTable to me (http://kodowa.com)

I had forgotten about lighttable, thanks. I'll need to revisit it.

What does kodowa do? I really don't understand.

Here's a demo http://www.chris-granger.com/lighttable/ (2 clicks from the linked page)

I always wanted something like that !

I think I might try a light version as a weekend project sometime soon, cheers.

don't know of any such tools, so instead I perform the task manually. If I want to better understand how some feature works, I write down in a text file what methods are involved, what method calls which etc. That way at any given time I have to answer fairly simple questions, but the accumulated info in the notes can then be used for more sophisticated reasoning. In other words, analysis :)

That's exactly what I find myself doing, but: I don't like paper As things change I find myself rewriting the same stuff over and over It feels like it should be easier/more effective if automated

yep, that's why my comment mentioned using a text file for notes, not paper. I usually edit it using IDE rather than text editor to take advantage of the convenient cut-paste a line of text with CTRL-X / CTRL-V without highlighting by mouse.

Then again, my nestgrid project http://www.nestgrid.org (the desktop version) is IMHO, among other things. the future of this sort of notes taking but I have not yet gotten around to using it "in production".