|Date||22 January 2012, 1300 CET|
|Participants||Peter Gorm Larsen, John Fitzgerald, Marcel Verhoef, Nick Battle, Shin Sahara, Nico Plat, Hiroshi Sako, Kenneth Lausdahl, Joey Coleman. Received appologies from Ken Pierce.|
The action item list is maintained as a tracker on SourceForge.
VDMJ: A few bugs fixed this period. A parser error was fixed that now disallows empty blocks; fixes were made to iota and exists1 to allow them to work correctly with type binds (for finite types); a small change to the type checker was made to better handle the type of blocks that “return nil”; and a larger change was added to perform better pattern matching, especially for cases where a pattern correlates the value of a variable from two or more place in the pattern.
IDE + AST: A few minor bugs has been fixed. One of the bugs was with the working directory not being set, this was introduced when fixing a mac debugger problem. The new IDE with the new AST is also progressing, all features are available except combinatorial testing and the quick interpreter. There is a few minor bugs left in the type checker which was introduced when a new location search engine was added to the IDE, this engine handles the editor outline sync and will also open for the implementation of more advanced completion.
Overture 2.0.0 has clearly been delayed; integrating the new AST into the interpreter is not yet finished.
A release plan for future features in Overture is pending some feedback on what future features we wish to add, and who is to add them. IHA is holding a brainstorming meeting in the coming week to work out possible local effort.
John Fitzgerald: The Strategic Research Agenda is due to be reviewed this meeting. Everyone is asked to read the relevant pages prior to the meeting. We well have an interactive discussion session dedicated to its contents. If you have concrete suggestions or comments to the SRA, please write them down here.
NB: Reading the second bullet on the main SRA page, it seems to say that the strategy is to treat Overture as an experimental platform (ie. not for stable industrial use), while VDMTools is the preferred industrial platform. Is that the intended meaning? It seems at odds with the description of the tools thread at the bottom of the page. The SEN paper we wrote a few years ago talks about building a platform to integrate VDM tools, but we don’t talk about tool integration much these days. I always thought that meant technical integration (at one time, you could select your VDM interpreter from a dropdown in Overture), but perhaps now it means that Overture is an umbrella term for various related VDM tools? Can we clear this up?
JSF: This reflects where we were a few years ago, before DESTECS and COMPASS got going. I think we were trying to say that we were targeting our research activity on making Overture an attractive platform for interesting extensions (e.g. in verification, testing, domain-specific modelling) that woudl arise primarily through research projects. If we intend Overture to be developed towards industry-readiness, is that part of the strategic research agenda?
NB: Also in the second bullet, it specifically mentions OO models and distributed/RT aspects as strategic. Does that mean non-OO VDM specification is deprecated in the strategy, or are we just responding to the current DESTECS type projects that are OO/RT. Does this mean the OO/RT aspects of the semantics thread have greater urgency?
JSF: I don’t think that the emphasis on OO/RT is especially appropriate now - it may have been more of a leading concern a couple of years ago. One observation from DESTECS is that VDM-SL definitely has a place.
MV: Removed several typos from the main page. It could use a simple graphic that visualise the main strands. More appealing then just text, in any case.
MV: Semantics sub-page: no reference to interest into probabilistics while this seems a logical next step. Furthermore, the link to (and the role of) the LB is missing. Who’s in control of the language?
JSF: If we want to make stochastics part of our agenda, that’s fine. It may be worth a workshop that looks at appropriate directions to take. In my experience, this needs to be thought through very carefully indeed, as it can become a semantic minefield. You need to be very precise about what you want stochastic analysis for, and what kinds of analysis you wish to do. For me, it makes most sense in the embedded systems domain.
See Planned Publications.
John Fitzgerald: Consider moving future NMs to Skype (text only, not video or audio in general). This has some advantage of making screen sharing possible for demos etc. For many of us, skype has become the norm for NMs in other projects as it appears to be more reliable for users on Macs. Against this, we do not have skype ids for all the typical Overture meeting participants.
All: Agreed, next NM will be on Skype.
John Fitzgerald: Consider moving the NM action list from SF to this Wiki. Upside: it’s one less window to have to have open in the NM. Downside: maybe we lose some tracking facilities (like all those useful emails that we receive when they change :-))
All: We keep the procedure as-is, just using a deep-link from the minutes to the SF action list. If action holders report progress prior to the NM then any discussions can be easy to edit during the NM.
Net Meeting 60: February 26th 2012, 1300 CET.