Wednesday, March 28, 2012

EclipseCon 2012 - I have had my share

I am at EclipseCon 2012, this year at Reston (VA). I have just finished my parts and can now relax and take it easy for the remainder of the conference.

Yesterday I gave a a tutorial on "DSLs for Xtext Developers" together with Sebastian. Around 50 participants learnded how to wirte their own DSLs that run on the JVM and integrate tightly with Java. Looking at the feedback so far, this was a real success, even thought it was a whole lot of stuff. Have a look at the slides:





Today I talked about "A Fresh Look at graphical Editing". In comparison to the talk at EclipseCon Europe, I changed the focus a bit and spent more time to make clear, why most graphical model editors really suck. This time I seem to have made my point, at least noone has asked me so far to convert the Generic Graph View framework into a model editor. The new features, like multitouch gestures and discovery diagrams went fine, too. Here are the slides



3 comments:

Juha-Pekka Tolvanen said...

Hi Jan,

Interesting slides and claims on graphical editing. For me they seem to dealing more with the challenges on getting satisfying modeling experience on Eclipse platform rather than (graphical) modeling in general. Nevertheless you also started mythbusting by claiming that picture is not worth of 1000 words. Please let me disagree.

Obviously you can take a poor graphical language (UML for metamodeling in your slides) and use that with better textual language (or do vice versa)… A better alternative would be making a comparison between proper graphical languages and proper textual languages. Last year xText was presented at Language Workbench Challenge to show basics but this year there was a complete domain to tackle. Could you implement the 2012 challenge with textual languages and then we could compare the graphical and textual approaches (see LWC2012 for details). This would also allow people to compare tools too.

- Juha-Pekka

Jan Köhnlein said...

Hi Juha-Pekka,

of course my (graphical) slides don't say more than I was sying in my talk either ;-)

My main point is that a good diagram is more targeted to humans (making them understand a complex model) than to computers (providing a complete semantic model). Thus, it shows a comprehensible subset/projection of a more complex model only. In addition it carries a lot information that is not necessarily relevant to the semantic model itself, like proximity of elements, placing things in the center, additional graphical emphasis etc.

Many graphical tools I know fail in creating such human centric diagrams. They focus so much on modifying the underlying semantic model within the diagram that they forget the importance of an agreeable design and good user experience. Or, even though they call themselves graphical, the user spends most of the time using non-graphical UI (trees, nested dialogs, tables...), e.g. to choose projections, provide fine-grained properties etc.

Unfortunately, my colleagues and me don't have the time to participate in every years LWC. Maybe we can do it the other way around and you present us a diagram that shows a really big amount of information and is easy to understand at the same time?

- Jan

Juha-Pekka Tolvanen said...

Hi Jan,
> of course my (graphical) slides don't say more than I was sying in my talk either ;-)

and for that reason your presentation could be provided as plain text ;-)

Jokes aside, it is indeed hard to add modeling capabilities to tools which were not built for that purpose (my old blog post on that). Modeling tools that use IDE philosophy for modeling capabilities end up having small canvas surrounded by “tools” that are needed for modeling (click element in canvas, open and browse it then in tree view to see properties which are finally available for editing somewhere else in the “tool”. And if editing is something else than text then the process starts again…as you said then “model editors really suck”).
It would be favor for the community that there would be also xText version available – especially since this year there was a lack of textual language workbenches. That was pity since the challenge included both specifying structures as well as behavior. If you don’t have resources to make the LWC 2012 challenge perhaps you can propose then a textual presentation for the graphical diagram version of the task? See the first P&I Diagram at pdf showing also other intgrated DSL.

What textual presentation would be better for that?
I agree that it is small – being my house – but we can naturally extend that for large office etc. buildings too. Ready for challenge?

Juha-Pekka