Poll: Open up reverse engineering

[danm]

First, EA is the most amazing tool I've come across in my 30 years of software development. I grew up on UML with Rose and EA is simply light years beyond, in functionality, usability and stability. I've been spending a huge amount of time with it and still haven't gotten my hands fully around it. There is practically nothing that isn't open for the user to tweak or modify, except one - reverse engineering.

Looking at the product that's the only glaring omission I can see - everything else can be modified in one way or another. Reverse engineering is an important final step that makes the whole process work. Unfortunately for me, I'm working in a non supported language, so I'm missing that final bit.

I can envision the following

• Minimally, an AddIn interface for reverse engineering. That should be relatively easy to support. For a new language you might be able to get this to work if you do it all yourself in an AddIn, but there seems to be some hooks missing as you can import, but not synchronize.
  
• Ideally there would be a new template mode, with similiar syntax to the existing templates.
  
• Or if nothing else, an MDA addin with support for this. MDA only allows reverse engineering for supported languages.
  

Poll - who else would like to see reverse engineering opened up?

[Uffe]

Gets my vote.

/Uffe

[Paolo F Cantoni]

And mine, FWIW - which doesn't seem to be much these days

Paolo

[«Midnight»]

Mine too, which is likely worth about as much as Paolo's. [But who knows, they might still add up.]

David

[Thomas Mercer-Hursh]

Add me in here as well, in spades.

[thomaskilian]

Would be nice, but we are still waiting for simpler things to happen.

[dbax]

I agree with Thomas' statement but I do give my vote for this.

[danm]

Hi Thomas,
  What other things are you waiting for?

Would be nice, but we are still waiting for simpler things to happen.

[«Midnight»]

Hi Thomas,
   What other things are you waiting for?

That's a question with a lot of answers - and no small number of opinions - from a lot of people.

Search through the forum, particularly this section. You'll find a lot of stuff.

Remember that EA is becoming a big product. Extension is occurring on a wide front, but still cannot satisfy all the needs expressed. Also, there is still a lot of 'homework' to repair or replace legacy functionality. Some is broken, some is getting obsolete, and some could simply use improvement. A lot of the core (and legacy) stuff isn't as 'glamorous' as nifty new functions, but would make the overall product more stable, etc.

On the whole things are moving forwards, and well. Still, I agree (very strongly) with Thomas et al, there is a lot of core stuff that needs improvement.

Just 0.02 CAD from the Frozen North.

David

[danm]

Sure, I understand. But I guess my point is that I expect the tool to work 'well enough' out of the box, and that it's more important to allow you to change the behavior, rather than Sparxs providing that.

For example, the code gen could be more capable. But instead of having Sparx do that for me out of the box, I'm happy to work the templates. If they're not powerful enough, hey Sparx gives me access to an AddIn. If all of that isn't good enough I can fully roll my own via an Add-In. I'm impressed with their 'levels' of tweakability.

Does that make sense? Reverse engineering is locked out, if it was opened, even in a minimal way, it gives me more power than enhancing existing functionality. You'll never get the existing stuff to work perfectly for everybody, better to let them tweak.

I guess the part I'm ignorant about is the broken functionality. I've trolled the boards quite a bit, and what I've seen so far seems rather minor, which is why I've asked. In my small experience the tool already performs wonderfully.

[thomaskilian]

danm,
we (old-timers) have been through quite some stages of enthusiasm/sarcasm/phantasm/... with EA. We still stick to it as it's the best out of a lot bad ones. Read Paolos (and if you dare bruce's) posts. There are so many tiny things. One would be an orthogonal UI instead of EAUI. Multiple select (anyone going for that). Multiple windows for diagrams/project browser. Copy/paste. And, and, and of course configurable reverse engineering. To put it simple: Abandon hope all ye who enter here (a bit too dramatically since Sparx is  doing a much, much better job than any other SW company I've seen). A lot of things are simply hard wired. Basically wrong by design. Since we're waiting for the basics for so long you need to be very enduring if you want to wait for a flexible RE. Honestly, I think the macro stuff for the FE is wasted effort. Automation can do much more. I'd rather like to see automation extended/improved. E.g. the cache in EA is worth nothing. You see the SQL server soooo busy when EA is parsing the tree.

Well, after all it's not too bad. I still live in hope. But don't ask for what...

[danm]

Thomas - I hear ya! I've read up on at least Paolo's missives ...

My small perspective, I last made a concerted effort with MDA back with Rose, across a Windows and Unix system for a 5 mil line monster legacy system, with a team of hundreds spread across the globe. Nightmares! We got it to mostly work though, but I learned to truly hate Rose, and IBM Architect ... forget it.

Now I'm on a smaller, new project and found EA - Eureka! Frankly, EA's sins are so small in comparison, as you say, I just don't care. Maybe a few years from now I'll be a jaded EA expert though ...  

Yes the scripts are crude - but that's the point, simple but very easy to change. OK if you want the PhD make an Add-In, but thank your stars you have the add-in capability so you can get the PhD - if you want!


"Multiple select. Multiple windows for diagrams/project browser. Copy/paste."

All good - but little to me frankly. After the horrors of my old system these are peanuts.