TEI Technical Council Minutes 2015-07
Teleconference 2015-07-28 14:00 UTC
Present:
- Syd Bauman (SB)
- Fabio Ciotti (FC)
- Hugh Cayless (HC)
- Martin Holmes (MH)
- Peter Stadler (PWS)
- Raff Viglianti (RV)
- James Cummings @ 14:23 (JC)
- Paul Schaffner @ 14:07 (PS)
Apologies:
- Stefanie Gehrke (SG)
- Lou Burnard (LB)
Agenda
What to do about Sourceforge
-
Reasons for leaving: wobbly infrastructure, shady practices, technology unlikely to improve. Are there good reasons to stay?
-
should we move to Git / GitHub? Are there alternatives we want to explore?
pros | cons |
---|---|
Better prospects for stability. | Migration process. |
Stylesheets and Simple are already there. | Possible loss of important history. |
Better forking/merging. | Loss of sequential rev numbers. |
Slightly better integration of ticketing and commits. | Slightly higher learning curve for new users. |
Pull requests / opportunity for better community engagement | Need to rewrite/update a ton of documentation and links. |
-
what needs to be replicated:
-
Keep presence on SF to avoid malware take-over. Possibly keep automatic mirror from GitHub
-
repository / software
-
issue tracking
-
file downloads (move to tei-c?)
- Files can be attached to GitHub releases (tags)
-
migration of issues: could be a tool: https://github.com/cmungall/gosf2github appears to be feasible for the migration of issues
-
-
The difficult thing is that the creator might get lost and the user using the API will be the new “creator”.
-
TravisCI in addition to / instead of Jenkins?
-
Timeline if we do move
-
Action on SM to produce a spreadsheet with info from tc documents that will need to be updated if we move to GH.
-
Action on MH to do an inventory of the SF site and produce a report on what’s there.
-
Action on HC to put a self-updating copy of the SF repo in github’s TEI-C account for safety’s sake and for testing.
-
Action on RV to look into the process of migrating issues from SF to GH.
-
Action on JC to investigate users on SF account and find out if they have GH accounts.
Fall F2F (HC)
- Sun 10-25 / Tue 10-27 in Lyon
- will be getting meeting room from U on Mon 26 and Tue 27
- need to find meeting room for Sun 25 — one not available, the other €900
- WiFi is a requirement, projection is a desirable
- Action on HC to look for “work space” or “share space” in addition to hotel meeting rooms
<s> inside <s> bug (MH)
- MH proposes returning to the old content model and implementing the restriction in Schematron.
- Council agrees: YES!
- Action on MH to raise a separate “<s> as child of <s>?” ticket
<xenoData> (SB)
-
how do we allow encoder to say “about the source” vs “about this TEI file”
-
should content be
-
well-formed XML-only, or
-
any well-balanced XML, or
-
any (well-balanced) XML or text() ? (The latter to allow “binary” formats like JSON)
- JC: I think we should make no rules about this other than that the document needs to be well-formed. (e.g. it can have text, CDATA sections, xml etc.)
-
-
The consensus is that anything (XML or text) should be allowed (of course it will need to be correctly escaped or encoded.) Stating the obvious: the result must be a well formed XML document.
-
HC: The issue of whether there should be a way to characterize what is in <xenoData> should be a separate ticket (@mimetype?).
-
FC: in Council Mail List I suggested to adopt something like @MDTYPE in METS
MEETING CLOSED AT THIS POINT.
- SB, MH, JC, FC remained on the call and discussed the nature of @rend… because they are gluttons for punishment. MH and JC agree that order isn’t significant (by default at least), and SB thinks it should be, but is wrong. ;-) To clarify that a little — MH and JC agree that the TEI should require that order not be significant, SB thinks that whether order is significant or not is up to the encoder / project / ODD-writer.
Their extraordinary meeting ending at 16:09Z.
-
default values (SB)
-
SB, LB, SR suggestion — make semantic suggestion rather than syntactic requirement
-
MH suggestion — make part of att.deprecated and actually remove
-