First of all, Bob and Tom thank you so much for discussing and sharing your opinions in the mailing list!
I greatly appreciate it, and all you write is very precious for me. Even though I am doing research in software engineering, it's very difficult for us to have the chance of getting feedback from "real" developers: - If you work in the research group of an industrial company, then you have the products of your company as a good case study for your research and you can get feedback from your colleagues developers; - else, if you work in an university research group that has contacts with some companies, then you can ask these companies to validate your ideas; - for all the other researchers (that's my case), it is very difficult to have good feedback on your ideas, other than that given by people in academia. So, thank you very much again. What I will do is to consider everything you tell me, and to share all the data that you want to make public with all the researchers who want to use them. In a real "free software" fashion. And, of course, results and tools will be free. On 26 May 2010 22:47, Tom Morris <[email protected]> wrote: > On Wed, May 26, 2010 at 3:57 PM, Bob Tarling <[email protected]> wrote: > >> I would think such a tool should preferably operate without any >> developer input but could then be tuned. > > My impression was that Alberto's questionaire was to establish a > ground truth of expert opinion against which to measure the results of > his tool(s). This is something that would be done during the > development/evaluation phase of the tool, not something that you'd do > as a matter of course if he were able to come up with suitable > automatic algorithms. You are both right: As Tom is saying, I need the ground truth of expert opinion to test my tools. These tools, as also suggested by Bob, will also use information from the versioning system to find the most changed classes. But not only this information. I would prefer not to give you now all the details about the information that we want to use for suggesting the "most important" classes. Otherwise, I could influence your answers in the questionnaire. But the general idea is to automatically derive some "metrics" from the source code, from the versioning system, etc. to find those classes. And then show them to the developers. > >> When viewing the results of the above the seasoned developwe may spot >> anomalies and then choose to remove items that are not relevant (for >> argouml that would almost certainly be the critics). I found Bob's idea of allowing a "human tuning" on the answers of the tool very interesting. Such a feature would be difficult to "validate", but knowing that you would find it important is already a good starting point for me ;) > > Except for a developer who wanted to extend the critics to cover UML > 1.4 or 2.x or modify the critic infrastructure to operate differently. > That was my point about importance being context sensitive. I agree > that for developers focused on other areas, critics could be largely > irrelevant (although the general knowledge that ArgoUML has multiple > threads of execution is probably important context for all > developers). Tom, I am trying to figure out how to include this context aspect in the questionnaire, because I find it pretty relevant. In the meanwhile, I would say that, if we can find good methods to tell you what are the most important classes *in general*, probably these methods could be applied to subparts of the system. For example, if we discover that the number of lines of code is a good way to find important classes in the whole system (I really hope not ;) ), I would expect that the same approach could work if you limit yourself to the classes in the GUI subsystem. > > BTW Alberto, I'd avoid the term "hot spots" and stick to "important > classes" or something similar. For me, hotspots is something I > asssociate with performance profiling (which may also be why Bob was > referencing code coverage utilities). Ok, sure. I will keep "important classes". Probably, in the questionnaire, I will try to define better what I would consider "important classes". Alberto ------------------------------------------------------ http://argouml.tigris.org/ds/viewMessage.do?dsForumId=450&dsMessageId=2613776 To unsubscribe from this discussion, e-mail: [[email protected]]. To be allowed to post to the list contact the mailing list moderator, email: [[email protected]]
