[Om3] Dates for second/third OM3 F2F Meeting

[Lars Hellström]

Michael Kohlhase skrev 2014-05-26 09.13:
> Dear all,
>
> it seems that the doodle has spoken. We have two dates, where almost all
> can be there
>
> - Mo, Jun 2. 14:00-15:00 (without David)
> - Thu, Jun 5. 17:00-18:00 (without James).

Any reason not to use the earlier 13.00-15.00 timeslot on that day which has 
identical reservations? Maybe off-line communication with clange (whom I do 
not see in the doodle)?

> Therefore I propose that we use both of the dates and proceed.
>
> Please save both dates in your calendar. I will send around an agenda
> nearer to the meeting.
> Please also keep up the good discussion on the TRAC.

Does that mean you'll be doing stuff there now? ;-) I mean, we[*] can all 
add comments to items, but in most of them (I believe) you're the only one 
that change any of the metadata (which in many cases is confusing rather 
than helpful).

[*] Well, James said (more or less) that he needs to renew his password. And 
Chris I don't know if he has a user to log in as yet. (David I just noticed 
has a user "david".) But once we've logged in, we can only add comments (and 
new tickets).


On the matter of metadata, here is one idea I've toyed with regarding a new 
set of Milestones: make them "If we put this in as a feature, then what 
would be the appropriate version number for labelling the product?" Some 
distinguishable cases should be

omstd 2.0.1: Exactly the same content as in the current standard, but with 
the obvious errors corrected; only pure bugfixes.

omstd 2.1: Only minor technical changes (at least of existing parts of the 
standard), but possibly larger changes in presentation. Working out in 
detail those aspects that are currently merely sketched would fall here.

omstd 2.5: Half-way to OM3; there may be new, important mechanisms 
introduced, but only in backwards-compatible ways.

omstd 3.0: New major version. We're allowed to break backwards-compatibility 
(but should of course only do so if we have a darn good reason for it).

An "OM2 standard, second edition" might mean any one of 2.0.1, 2.1, and 2.5; 
in my experience, a new product will be released whenever (i) it is ready 
and (ii) there seems to be some point in releasing it.[**] Even if minor, a 
2.0.1 would be an improvement over the current omstd2.0. But if we get a 2.1 
ready while we're at it, then it could just as well become the "second edition".

[**] Well, I'm omitting the crucial requirement (iii) someone bothers to 
wrap it up, which is often the real blocker.


Some examples of how I believe items should fit into the above scheme:

omstd2.0.1:
#9 (Update to new URI spec);
#83 (Default CDBase is only specified quite implicitly);
#104 (OMR examples false);
#106 (Specify a MIME type for OpenMath);
#107 (Binary encoding is bigendian);
#108 (typo in OMSTD2.0 chapter);
#114 (Endianness of OMF);
#138 (CD's CDBase declaration is mandatory);

omstd2.1:
MathML encoding (big, but not a major change to anything previously 
established);
#1 (Definition of alpha conversion underspecified);
#75 (general formulation issues);
#109 (Roles and Complience);
#126 (CD versioning probably not compatible with CD URIs);
#127 (clarify specification of CD URIs and CDURL for linked data);
#128 (Make CDSignatures or Signature use cdbase);
#131 (Disambiguate CDGroup/CDName);
#134 (Specify or forbid whitespace normalisation);
#143 (Variables not bound in head of binding);
#146 (Bit numbers start at what?);
#151 (STS documentation is unclear about binders);

omstd2.5:
#28 (New Roles for Types?);
#32 (FMP type=alias);
#72 (Title as a child of CD as well);
#129 (Introduce computational FMPs);
#136 (n-ary OMBIND);
#144 (Add Notation Definitions to OpenMath);
#145 (Allow OMOBJ in OMCD texts);

omstd3.0:
#11 (Mixed content for textual bits in CD descriptions);
#125 (Drop FMP/@kind) -- it's minor in practice, but would turn valid CDs 
invalid;
#147 (Quoting of OMOBJs as part of an OMOBJ);

As it turns out, I ended up going through the whole Trac (but skipped a lot 
of items). And I may have misinterpreted some things. Still, the above 
should be useful.

Lars Hellström


>
> On 19.5.14 12:36, Michael Kohlhase wrote:
>> Dear all,
>>
>> I had the homework for making a doodle poll, it is at
>> http://doodle.com/m2x37caccpins3gs
>>
>> Michael