Umple

Just in case I replied only to Tim. ~ is syntactic sugar for alphanumeric. -- aforward

The issue identified in B2b (losing context) I think is a big deal and something that would need to be seriously addressed before moving away from status quo. I am all for helper classes (eg StringTracer), but the helper would need to truly be an ... [More] abstraction of the problem domain and it appears like tracers are a little "powerful" to fully delegate to one "trace" call (because you lose the context of the caller -- and abstracting that out I think would really constrain your ability to trace any "class"). [Less]

just a brief comment on option C. this already what we are doing for StringTracer, I would just be proposing to take that design one step futher and do if for each tracer...(just wanted to be clear :))

OK, Here is the last message in the sequence, and here are my comments: We need to set critieria on which to judge the varous proposed solutions. Here they are (not necessarily in order): 1. Run time efficiency 2. Flexibility to adapt the language along several dimensions a) Multiple language support (java, cpp, PhP, etc)

---------- Forwarded message ---------- To: Geoffrey Guest <[email protected]> Cc: Timothy Lethbridge <[email protected]> I see your point. Tim for sure will have a comment on this. I brainstormed a solution which uses generator to produce MOTL class. First, check the implementation of LTTNG tracer in Cpp. mainly check the

---------- Forwarded message ---------- To: Hamoud Aljamaan <[email protected]>, Timothy Lethbridge <[email protected]>       What needs to be added to JET templates is the MOTL class itself. yup, i originally had that hope, except that's not how the jet files work. They work on umple classes, the file that they output to is

Unfortunately today's email readers are crap and don't consistentlly annotate quotes: ---------- Forwarded message ---------- To: Geoffrey Guest <[email protected]> Cc: Timothy Lethbridge <[email protected]> Hi Hamoud,This is the direction I originally wanted to go, until I got to trying to implement it.

I will forward the rest of this chain of thought and then comment. ---------- Forwarded message ---------- To: Hamoud Aljamaan <[email protected]>, Timothy Lethbridge <[email protected]> Hi Hamoud,This is the direction I originally wanted to go, until I got to trying to implement it.

Hi all, There has been an ongoing conversation between Hamoud and Geoffrey about refactoring how the tracing subsystem works. I am going to forward these to umple-dev, and inject my comments at the same time. I would appreciate if the most experienced team members including Andrew Forward would inject any thoughts.

Hi, I got back from my errand, and the build is now fixed. Timothy C. Lethbridge, PhD, P.Eng., I.S.P., CSDP Professor of Software Engineering and Computer Science / Professeur Titulaire de gￜnie logiciel et d'informatique Facultￜ de genie / Faculty of Engineering University of Ottawa / Universitￜ d'Ottawa