Meeting minutes
Agenda Review
Action Items
<addison> https://
addison: r12a, should I do your bidi related action as part of my next editing round of string meta?
<addison> action-975?
<trackbot> action-975: Addison Phillips to Update wiki on i18n "considerations" sections and reply to sec/ping thanking them -- due 2020-11-19 -- OPEN
<addison> https://
addison: did work on this: updated wiki ...
… please have a look, comments are welcome
<addison> close action-975
<trackbot> Closed action-975.
addison: will close action
action-1017?
<trackbot> action-1017: Addison Phillips to Send wpwg list of specs that implement lang/dir -- due 2021-04-22 -- OPEN
pending
<addison> action-1024?
<trackbot> action-1024: Addison Phillips to Review floating times article for final publication -- due 2021-05-13 -- OPEN
addison: done, made some edits
… merged back in
<addison> https://
addison: does the group want to review this?
<addison> close action-1024
<trackbot> Closed action-1024.
addison: we did a wide review in the past
… I made only small edits
(no disagreement for publishing the doc, but table the doc for now)
<addison> action-1030?
<trackbot> action-1030: Atsushi Shimono to Ensure that a summary of the ruby issue in english is forwarded to the working group -- due 2021-06-03 -- OPEN
atsushi: sent out on Monday
<addison> close action-1030
<trackbot> Closed action-1030.
<addison> action-1031?
<trackbot> action-1031: Addison Phillips to Document the state of payment-request -- due 2021-06-03 -- OPEN
addison: pending, will work on it today
Info Share
xfq: W3C process will have a new version this year
… a new "note" track will be added
<xfq> https://
xfq: is separate from REC track, to avoid ambiguity
… above link shows the draft of the process
<atsushi> all changes
addison: we are rechartering, so everything will be under process 2021?
xfq: all specs will switch to new process once it is published
… patent policy is depending on the charter
addison: those are often linked
… there is probably not much impact on us
… a new version of respec will incooprate the stuff, I guess
RADAR Review
xfq: yes, should be transparent to all groups
<addison> https://
addison: some new items
(addison going through the list)
atsushi: happy to make a review
… resource timing and user timing
addison: we will discuss this in 1 or 2 weeks
bert: interested to review dx profile
"dx profile negotiation" (data set negotation)
"Content Negotiation by Profile (dx-prof-conneg)"
bert: will have a look
manifest and localization
<addison> https://
<addison> https://
<xfq> Open tracker issues for WAM: https://
addison: a search for what the best patterns should be, to create a manifest for a web based app
… two common models are debated:
… a single file, contains different localization
… or: multiple manifest files, one per localization
… which do we recommend?
… what should be best practices for a space like this?
xfq: a third model would be:
… a file that contains not a whole manifest file, but some localized strings
addison: a resource file attached to the manifest
xfq: correct
… each has pros and cons
… for web apps, they need to fetch multiple files, there is a cost for fetching
… there may be potential duplications
… with a single file, it is difficult to modify
… if there are a lot of locales, the file is large, not easy to manage
<JcK> +1
atsushi: separation is easier for distributed work
… quite easier to have separated files
… of course post processing could be used to create one file
… some apps may set english as default language
… or if some language is lacking, a value could be copied by the processor from default language
<Zakim> fsasaki, you wanted to agree with atsuhi :)
<addison> felix: want to agree. in localization workflows, if you have monolingual files it's easier
addison: agree ... there are several topics in here ...
… one has to do how language negotiation can occur ...
e.g., how do you select the localized resource ...
… that is there the default value comes into play
… you also want to reduce the number of fetches ...
… if you have a localized resource file, you want to know what to get, based e.g. on language negotiation
… final consideration is the size
… if you do a lot of languages, you can have a huge manifest file ...
… having dozens of locales would lead to a big file
jck: general idea of an index file
… and a default, and a bunch of URLs, doesn't that solve the problem?
addison: would solve the problem
… lang neg. is then on the client, not on the server
… there is a couple of different models
… some manifests go into a zip file which forms an app
… if you bundle all in zip file you can have everything inside
… you do not fetch things, you just go inside the directory and get what you want
… there was a directory structure with all kinds of stuff
… if you have a manifest, you want to fetch as less resources as possible
… that would then be a multilingual file, or a file with links, a list with what you need
jck: the latter has two fetches
addison: which is better than 50+
jck: and if you have thousands of resources, zip still needs a lot of time to retrieve
atsushi: android jdk takes such an approach
… they use the "separate" model for storing localized string
addison: yes, and then everything comes in the zip file
… you download everything and read it
… which is also convenient for localization processes, as felix mentioned
addison: how to move forward - should we contribute to the thread?
"best practices for manifest files"
xfq: the problem is different from string metadata problem
addison: it is related to that problem
… since you will need string metadata in such situations
xfq: correct, and in such a file you can still have strings with different languages and base direction
xfq: correct
r12a: people in that thread were starting to talk about many different topics
… some examples seem to assume that you always know the language of the document and the direction
… so this needs to be distinct from string metadata, and then show how things work together
Action: addison: create skeleton of "best practices for manifests"
<trackbot> Error creating an ACTION: could not connect to Tracker. Please mail <sysreq@w3.org> with details about what happened.
addison: once the skeleton is done, people can look at it for further work on the doc
… anything else we want to do on the discussion thread?
Action Items
action-1032?
<trackbot> action-1032: Richard Ishida to Propose text for html 4814 (datalist find/string search) -- due 2021-06-03 -- OPEN
addison: how about the action on bidi keywords?
r12a: the glosarry is the place this should go to
<r12a> https://
r12a: we have something in the glossar, not sure if that is what you have in mind
<r12a> https://
addison: would expect this together with some explanatory text
… a definition with usage
r12a: could go into an article or document
r12a: please do the edits you had in mind
addison: ok, will do
Info Share
<r12a> https://
r12a: added more stuff to glossary, please have a look
<r12a> https://
r12a: looked at char mod, character essentials etc., this is growing, keep an eye on it
r12a: this is also a reply to one of atsushi's github pull requests
… mentions gap analysis pipeline
… tells you how to create a gap analysis document, via github
… and describes labels to use
… which is relevant for the pipeline
<addison> (notes that LDML is UTR35 in bibxml I believe)
r12a: if you do gap analysis, please have a look at this
… also, we had a discussion about 2021 process , cf. the note track (mentioned by atsushi)
… we can also have "super notes"
… means: you have wide review and s.t. else
… and show that people reviewed it
… it then it gets the "w3c endorsed note" status
… relevant for various docs, e.g. "string meta"
… can be applied to existing docs, but needs evidence that there has been a review
addison: so sort of like a REC, but without conformance etc.
… a bit like BCP in IETF
r12a: notes can be all kinds of shape in W3C
<xfq> https://
r12a: it might be, that for our docs would be relevant for this status
… name of the new document category is still under discussion
addison: would such a document go through the same gates for publication?
<xfq> https://
r12a: not sure yet
<JcK> "Statements" are a step down the slippery slope from IETF BCPs which can contain all of the categories you (Addison) mentioned plus nearly pure rants
xfq: it seems that we can do editorial changes, but with non-editorial changes, there is some processes
xfq: editorial means: fixing links, markup etc.
atsushi: note is similar to working draft
… wide review, AC review apply
addison: a bit like REC track, but not normative and without implementation requirements
… might help us with some challenges we had in the past
… e.g. with regards to references to char mod norm
jck: useful for many cases
addison: is there a requirement during rechartering to list work items with this target?
r12a: no
… process is not in place yet
addison: do you expect that to be in process 2021?
r12a: yes
<JcK> But problematic is not applied carefully and used very conservatively
<addison> agenda
no more infoshare
AOB?
adjourned