See also: IRC log
<Ashok> Scribenick: Ashok
<scribe> Scribe: Ashok Malhotra
Noah: Goals are to decide what we can do to move ahead or decide not to move ahead
... Tim the TPAC committee is waiting for you to tell them if you want to talk about it
Tim: Can we agree to define the charter of the group here
... we have started a task force ... Raman asked for it
... asked Norm to join ... said yes
<noah> I think what Larry suggests is reasonable IF we can convince ourselves that this could be a realistic effort, that we have a sense of what timeframe for impact, etc. Then we can think about soliciting participation.
Raman: We should not call it a TAG task force
<noah> Also, I think this can't be a volunteer effort -- I think the point is that Tim, with our help, will pick a small, balanced set of experts to see if some new avenue can be discovered.
Raman: the charter will determine who participates and who participates will determine how well the task force does
Tim: Agree charter is most important ... will decide who joins and how successful we are
<jar_> coordination?
Tim: Divergence of XML and HTML
LM: My understanding of this isuue ahs been hampered beacuse there are no participants from XML community
... we need XML usecases etc.
Raman: We need people who are most affected
... Tim, you had a list on Tuesday
<noah> Noodling on charter: Goal is to develop specifically technical proposals and/or best practices that would increase the ability to 1) use XML tools to create and manipulate HTML and 2) to include XML fragments in HTML and 3) to include HTML fragments in XML.
Tim: --subset of HTML5 needed for simpler devices
- Impossible derive a RESTful API comapred to XForms ... cannot tell waht data it transmits
<noah> Noodling on charter: Goal is to develop specific technical proposals and/or best practices that would increase the ability to 1) use XML tools to create and manipulate HTML and 2) to include XML fragments in HTML and 3) to include HTML fragments in XML.
<noah> Noodling on charter: Goal is to develop specific technical proposals and/or best practices that would increase the ability to 1) use XML tools to create, manipulate and "round trip" HTML and 2) to include XML fragments in HTML and 3) to include HTML fragments in XML.
Tim: Also use XML tools to create HTML and load XML
<noah> Addition to the charter: The requirement is to do this in a way that will be widely deployed.
Noah: Whether to push well-formed should wait until task force is formed
Tim: What is important for XML folks is infoset and what is important for HTML folks is DOM
Raman: Make sure you deliver good DOMs via good markup
Noah: The charter should start with broad success goals
<Zakim> noah, you wanted to discuss charter proposal
Noah: start by saying the taskforce has to find something that widely accepted
LM: It is not a good idea to talk about XML community and HTML community
... if we are successful there will be a single community
<noah> What I said more specifically was: the charter should require a solution that will be widely deployed, for the stated purposes. Achieving that will require the task force to confront the very difficult technical challenges, e.g. that well-formedness is seen as conflicting with existing practice, is cumbersome for some, and so can't in general be required.
LM: some use cases such as EXI and DSig work for XML but not HTML
... perhaps increase scope of polyglot documents
Tim: Raman spoke about organizing the ultimate converence not a quick fix
<noah> To my 3 charter points, I might add: attempt to increase the set of documents that behave as polyglot/chameleon, I.e. that can be served with reasonably compatible semantics, DOM, etc. as either text/html or application/xhtml+xml
Raman: HTML5 spec is starting to add rules
... such as Table/Tbody
... Webkit will fix bad Table syntax
... and serialize correctly
... I think these kinds of rules will become common
... problems are with things like xml:id and id, xml:lang and lang
... leads to 2 worlds
Noah: XML community may need to move
Raman: Larry talks about tools ... HTML tools was IE
... XML tools are there and are used ... legacy
... You said creating HTML from XML is easy ... I don't think so
Noah: The XMLstuff is of value to some people ... they have deployed code ... that is why we had trouble with stuff like XML 1.1
... Many of the changes would change the value proposition for XML
... they would resist change
<Zakim> ht, you wanted to disagree with TV's analysis: there are two XML communities
<johnk> masinter: isn't the standard HTML <-> XML transformation the thing (formerly?) known as XHTML5?
HT: There are 2 XML communities ... people who use XML as XML for interprocess communication, not interested in human readers
... there is another community who care about human readers. The 2 communities use some of the same tooling but have different goals
<Zakim> masinter`, you wanted to talk about transformation gateway, ask if their's a value to standardizing that
LM: XML/HML transformation gateways ... differences in the DOMs
<jar_> wannabe-commutative diagrams (html / xml / source / dom)
<raman> agree with ht. but the interesting case for convergence are formats like rss or atom. So is Atom/RSS data or content? Today, the content particle in rss/atom is html -- and is carried as a bag of bytes --- not as something that can be parsed as xml.
<raman> agree with ht. but the interesting case for convergence are formats like rss or atom. So is Atom/RSS data or content? Today, the content particle in rss/atom is html -- and is carried as a bag of bytes --- not as something that can be parsed as xml.
LM: How to apply XML DSig to Webpage?
<johnk> http://wiki.whatwg.org/wiki/HTML_vs._XHTML
Tim: Including scripts?
LM: That's a kind of workflow where we do not know that the standards work together
<Zakim> noah, you wanted to ask how this is positioned relative to HTML5 specification development process
LM: use this kind of value-based reasoning to drive the taskforce so it is nor percieved as an academic exercise
Tim: The Polyglot investigation did address that
<jar_> lm: Let's identify the real value of using XML, e.g. signed content, and suggest that it would be awfully nice to get the same value for HTML
Tim: there are some constraints ... I'm happy to live with these constraints
... happy not to write scripts that use some DOM properties
... polyglot defines some rules and if you follow them everything just works.
JK: Sceptical of that claim
<Zakim> noah, you wanted to ask how this is positioned relative to HTML5 specification development process
Noah: The chairs will ask how this affects them logistically
... questions about process to complete specs, etc.
Tim: The task force will take the polyglot idea and take it further ... polyglot idea did not come from HTML5
... Do we feel about changing XML so the XML parser would use the default namespace from the MIMe type
HT: I talked to Liam ... this may be acceptable as a change to XML
Yves: Include XML fragment in HTML ... this is the only one that will have impact on HTML
Noah: Task force will decide what to recommend
Yves: Should the task force work on pussing XML fragments in HTML
<noah> NM: I think the task force should try to figure out which of these specific changes, in XML or HTML specifications, would likely be deployable and a win for the communities as a whole
HT: You edit the doc ... edit the SVG still works because using a HTML5 parser ... then you cut it out and it fails
Yves: You have a serialized DOM ... issue on the tools not a language problem
LM: View Source may have some options
<noah> 4 mins to go on this subject
Noah: Tim, do you have what you need re. talk at TPAC and deciding whether and how to charter the task force?
Tim: GOAL of task force: Enable and enhance mixed XML/HTML workflows
... We have a good story and we can go ahead
... need 6 more folks
... Steps moving forward: Draft charter, circulate to TAG, Circulate to candidate TF, Circulate to AB,
LM: We may also need an IG
HT: You may want a period of negotiation with the candidate TF
Break for 10 minutes
<DKA> http://edition.cnn.com/2010/TECH/mobile/10/20/html5.smartphone.apps/
<DKA> HTML5 is the future.
<DKA> ( apparently )
<noah> close ACTION-459
<trackbot> ACTION-459 Schedule F2F discussion of IRIbis status and Larry's proposal closed
Noah: HT wanted a f2f update from Larry
<noah> ACTION-410?
<trackbot> ACTION-410 -- Larry Masinter to let the TAG know whether and when the IRIEverywhere plan in HTML WG went as planned -- due 2010-11-01 -- OPEN
<trackbot> http://www.w3.org/2001/tag/group/track/actions/410
HT: The TAG should hear about the way in which you rearchitected the way in which the specs will work
... what happened? Did the idea succeed.
LM: There is a spec ... above
... there is a WG chartered to finish the spec ... we are moving forward
... There were 4 different things
<noah> Private conversation: DanA kindly agrees to integrate minutes for Tuesday
IRI's could occur in:
XML
HTML
OTHERIRI
FORMALIRI -3987
scribe: LEIRI had its own IRI and had an algorithm for transforming to FORMALIRI
... URI is a FORMAL URI but also lossy
HTML URI was not well defined ...
<DKA> [ Apropos to the privacy discusion: Google Engineer Builds Facebook Disconnect (TechCrunch): http://tcrn.ch/9zFyot ]
Added "PRESENTATIONS" below HTML
scribe: what you write on the side of a bus is not a sequence of codepoints
... impt to separate out the presentations from the sequence of unicode characters
LM: Deprecate the idea that there is a set of chars that are not allowed ... there is a syntax that determines what is allowed
IRI -> sequence of unicode parts
IRI components -> processing
<DKA> Image of whiteboard from XHTML-HTML taskforce discussion: http://www.w3.org/2001/tag/2010/10/xmlTaskforceWhiteboard.jpg
JR: There is a presentation, there are octets and then parts of the IRI
IRI is a sequence of codepoints
scribe: someone has to pick a character encoding
<ht> iso-8859-1, etc.
LM: The parsing happens on the IRI and from that you get a set of components and this is what you process
test
LM: What HTML was doing was not a syntactic item ... it was a processing step ... step can know about document encoding
... We now have a document with correct structure but lot of the details are still wrong
HT: The splitting into components is character independent
LM: There is a change proposal for HTML spec that points at this
Noah: How is HTML WG looking at this?
LM: Adam authored an alternate change proposal
... folks want this problem to go away
... 2 different versions of IDNA? how do we handle bi-di?
... for Bi-Di we need directional markers
... this is still up in the air
... trying to get critical mass
... There is a document ... still has bugs ... needs people to take a look
HT: This is good stuff. If the politics get sticky let us know.
LM: Politics is sticky
... Need to change registry to register IRI schemes
Tim: How do I need to change RDF code ... now it normalizes the URI
LM: We did not change URIs
Tim: If it is not URI it is hex encoded
LM: If you had a Chinese assertions the hex encoded URI will not work
Tim: This breaks 10 yrs worth of code
HT: The hex encoded URI will fail unless client libraries update
LM: It was a hard choice ... none of the choices were great ... I think this was the best choice
<noah> ACTION-410?
<trackbot> ACTION-410 -- Larry Masinter to let the TAG know whether and when the IRIEverywhere plan in HTML WG went as planned -- due 2010-11-01 -- OPEN
<trackbot> http://www.w3.org/2001/tag/group/track/actions/410
<noah> close ACTION-410
<trackbot> ACTION-410 Let the TAG know whether and when the IRIEverywhere plan in HTML WG went as planned closed
<jar_> ACTION jar to assess potential impact of IRI draft on RDF/XML, OWL, and Turtle
<trackbot> Created ACTION-487 - Assess potential impact of IRI draft on RDF/XML, OWL, and Turtle [on Jonathan Rees - due 2010-10-28].
<jar_> action-487 due 2011-12-01
<trackbot> ACTION-487 Assess potential impact of IRI draft on RDF/XML, OWL, and Turtle due date now 2011-12-01
<jar_> http://www.w3.org/2001/tag/2010/10/metadata-arch-outline.txt
Noah: We need to discuss about how to proceed on Metadata
jar: Just typed this is in last night ... goal is to prevent bad things from hapenning
... lots of metadata efforts
... is my list about right?
... Trying to put things into a framework
<DKA> Another metadata effort I have been involved with in the past: PRISM - http://www.idealliance.org/industry_resources/intelligent_content_informed_workflow/prism
Noah: Goal is to guide the community ...
jar: Options on how to add metadata ... how to choose
<DKA> Also - POWDER : http://www.w3.org/TR/powder-dr/
jar: If I get guidance i can produce draft by early spring
Tim: These are things we can do and how to do it ... where are the orange blobs [reference to diagram on whiteboard, blobs are research areas]?
... crises or opportunities?
... I need a heat map
<Zakim> masinter`, you wanted to talk about: bad things that can happen, other issues
<timbl> Threat level opportunity level
LM: Big space, stuff is moving
... preventing bad things has more urgency
... people choosing metadata vocabularies that are different to map between ... same data. different languages
<timbl> Threat: Lack of harmonization of metadata ontology
<timbl> --Larry
LM: Vocabulary divergence
... the other is when you have multiple sources of metadata and how to pick what you want
<timbl> Metadata services / indexes
<timbl> . . . SPARQL
LM: possible conflict between metadata sources ... metadata in lots of different places
... we are at risk of W3C exacerbating the situation
... The outline is fine
Noah: Probes on what LM wants to proceed
Tim: Orange blog on create/curate formats?
<Zakim> ht, you wanted to ask about metadata defined how?
ht: metadata for what?
... what's the domain?
jar: Metadata is data about digital objects
<Zakim> masinter`, you wanted to point out media annotation working group work
LM: There is a media annotations WG ... they have a document about metadata and metadata resolution etc.
... perhaps ask if their idea of metadata and metadata resolution matches ours
<ht> hst: "data about digital objects" is fine with me, against a background of focussing/foregrounding the Information Science community's view of what metadata/digital objects is/are
jar: Looks like a format ... for delivering they reference powder
... Easy answer is Semantic Web gives you an answer [JAR looking at these minutes has a feeling he said "RDF" not "Semantic Web"]
<Zakim> timbl, you wanted to say that at the moment of course the digital objects out there of great interest are [government] linked open datasets. It is very timely to encourage work on
<timbl> http://www.ckan.net/
Tim: UK Govt uses CKAN
Jar: You register things in it so that they can be found
Tim: Has an API ... not particularly RDF oriented ... cannot do Math
... Has a keywork query info
... Connect up people who are working on repositories with SPARQL code
<DKA> Just looking at http://www.w3.org/TR/mediaont-api-1.0/ (API for media annotation API) and thinking about in the context of Web Application APIs and also in the context of the metadata discussion...
Ashok: Should the TAG recommend "use RDF (and OWL) for metadata and SPARQL?
Tim explains 5 star rating for data
<noah> JAR: That's new in the outline: discussion of APIs to access metadata
<timbl> + to talk about Jeni Tennison's work
<johnk> http://code.google.com/p/rdfquery/ is Jeni Tennison's rdfquery work
Tim: Emphasize, that RDF & SPARQL for data and for metadata
... Jeni's work ... RDF repository with SPARQL access
<ht> http://www.legislation.gov.uk/
<DKA> Interesting blog post from Russell Beattie on microformats in HTML5: http://www.russellbeattie.com/blog/html5-microdata-thoughts-on-semantic-data-mobile-adaption-and-facebook-fbml
Tim: typical user does not understand/use SPARQL
<jar_> this is off topic
Tim: there is an API that generates SPARQL
<jar_> very interesting but it has nothing to do specifically with metadata
Noah: Pl. clarify what happens at client and what at the server
Tim: We need to say how to use the SPARQL
<Zakim> noah, you wanted to ask about going too far
Jar: How best to use the RDF stack is a goal and work in progress
... Here are the benefits you get if you use RDF and SPARQL
Noah: We could write good practice notes
jar: If someone is immersed in another metadata format we have to be more nuanced
... make properly qualified statements
<Zakim> masinter`, you wanted to argue that "metadata" is a different topic from "data"
<jar_> We lose credibility if our statement about the goodness of RDF is too broad.
LM: If everyone uses RDF but different vocabularies that would be bad. We want to say RDF is processable ... and use same vocabularies
<jar_> TBL said matching vocabularies is easy compared to getting the information formatted in the same way. JAR said that in his experience, format conversion is easy, relating semantics is very hard.
LM: If data formats are different then it may not be a big problem
... Some more immediate things we could do ...
Tim: What will the finding be ... Larry is pointing out a hotspot
LM: Two areas hapenning in W3C where we can help - microformats vs. RDFa and Media Annotations
Tim: Need some leadership in the area of metadata fo Linked Data
<noah> ACTION-282 Due 2011-04-01
<trackbot> ACTION-282 Draft a finding on metadata architecture. due date now 2011-04-01
LM: Personally, I would put this as higher priority than Persistence
jar: [persistence is higher priority for JAR]
<noah> John to integrate Wed. minutes
<noah> Jonathan to integrate Thurs. minutes
<jar_> TBL: Star rating for linked data.
<jar_> 1. on web
<jar_> 2. machine readable
<jar_> 3. non-proprietary
<jar_> 4. in rdf
<jar_> 5. is linked
<scribe> Scribe: Jonathan Rees
<scribe> Scribenick: jar_
<noah> How are we doing with scribing this afternoon?
Dan A introducing web app API topic
DKA: maybe aim for something could be useful to future people working in the space
<noah> Where's the threat, and where is there confusion?
DKA: what architectural principles might apply to API design?
... What other technologies are core building blocks, things that one is advised to build on?
... It would be beneficial to do some work in this area
masinter: Re APIs, there's a perspective that the web doesn't need versions.
... With content, you have ignore-unknown and so on
... But how do you do API evolution without versions?
... How big is the API, how do you decompose it. What happens when the language itself evolves?
... Are we talking only javascript, or also Java and so on?
... These are the kinds of questions I have in mind when we talk about architecture.
<Zakim> noah, you wanted to talk about composability?
noah: TAG should emphasize things that help the whole to come out right. Fit well compose with other things
... There is some of this in the Javascript world
... We could emphasize the things that have to do with the Web
<Zakim> masinter`, you wanted to ask some questions about versioning of APIs, granularity of APIs, data structures, evolution of JavaScript
<Zakim> timbl, you wanted to say Multi-lamguage APIs which the DOM aimed for are sub-optimal for tight binding of language t the data. 2) versioning? Programmers cope with bits not being
noah: What happens when there are multiple events/handlers any of which can fire (e.g.)
timbl: Back when, you were supposed to make an IDL, then bindings. What programmers look for now is a much closer binding of API and language (e.g. iterators). New pattern of higher-order functions.
... IDL may be too limited... short names are good, you can assume names are scoped [unlike IDL style]...
... instead of calling a method on something, just iterate over it... seems the way to go
<noah> I think IDL is motivated primarily when you want to support multiple languages, such as VBScript. I agree with Tim, getting JavaScript right, and benefiting from idiomatic constructs is the right way to go.
ashok: Touches on: How do you save state? How do you communicate with other apps? Storage? Authorization? We ought to coordinate this work
<noah> By the way, Tim pointing out the JQuery/Dojo idiom of chaining functions is a great example of using an overarching architectural approach to get composability.
dka: Re inter-webapp communication, look at Open Ajax Hub. Creates a secure environment for this.
... Between different frames
ashok: Example?
dka: 2 apps in different tabs, one wants to communicate with the other
noah: Restaurant app + map app mashup
dka: Ashok, you were talking about a local app in browser wanting to talk on an ODBC connection to a local database not in the browser? Nobody working on that
noah: What if we *don't* work on this?
dka: Privacy is bigger and more urgent compared to APIs
<noah> NM: I tried to say, what do we lose or what do we slow down if we work on this?
ht: EXT
jar: What about toolkits, if we were to work on this we might see what these efforts are and if they're happy
noah: Put off til spring, do privacy in meantime? Or what?
<DKA> action-461?
<trackbot> ACTION-461 -- Daniel Appelquist to draft "finding" on Web Apps API design -- due 2010-10-31 -- OPEN
<trackbot> http://www.w3.org/2001/tag/group/track/actions/461
dka: I'd like to talk to people at TPAC doing API design
<noah> . ACTION: Appelquist to solicit at TPAC perspectives on what TAG could/should do on APIs
dka: to get better understanding of how TAG work would be received, and get input
<noah> . ACTION: Appelquist to solicit at TPAC perspectives on what TAG could/should do on APIs Due: 2010-11-09
<noah> ACTION-461 Due 2010-12-31
<trackbot> ACTION-461 Draft "finding" on Web Apps API design due date now 2010-12-31
<noah> ACTION: Appelquist to solicit at TPAC perspectives on what TAG could/should do on APIs Due: 2010-11-09 [recorded in http://www.w3.org/2001/tag/2010/10/21-minutes#action01]
<trackbot> Created ACTION-488 - Solicit at TPAC perspectives on what TAG could/should do on APIs Due: 2010-11-09 [on Daniel Appelquist - due 2010-10-28].
ashok: We ought to produce something on evercookie thing quickly
dka: draw on recent posts to www-tag
... "privacy" is a big thing, maybe shift subject to user intent
<noah> NM: I like thinking about this first in terms of: there is a risk that software will be able correlate events that the user does not wish to have correlated. E.g. the same user who bought this house at a real estate browsed for bars at some search site.
masinter: [see above] Really clearing private information is really hard, you have to cleanse every server
... Someone might be confused about this
<noah> NM: We can explain that, among the reasons this causes concern, is that such correlation can sometimes allow discovery of information that the user might wish to keep private (I don't want the people who are selling me the house to know that I drink a lot.)
dka: The scope in an ASAP statement should be limited to evercookie, how standards community might respond
ashok: We could say: browsers ought to let you clear private information. That's tricky to say
<Yves> think of spam, link to remove from "spam lists" that are there to check the validity of emails... and the never-ending spam/anti-spam race. same issue with tracking and cookies
<Zakim> ht, you wanted to ask about the connection with JAR's perspective
ashok: ... what else could we soapbox about?
ht: Butting heads: You must be able to clear browser state / You can't clear browser state
... That you're trying to protect your own private
<Zakim> noah, you wanted to discussing referring expression
ht: The way to achieve desired effect is to institute accountability
... [...] is not a referring expression
noah: Given that we can't plug all the holes, should we try to plug any?
... We can say that there are other ways to look at the problem
... Can we settle that question?
... The steganography example (URI history list) is what pushed me over
raman: As we go mobile, the trend is to be able to easily share context... URIs communicated from browser to phone
<noah> NM: The key thing I said is that I think the TAG should start by settling the question of whether Evercookie just shows that the list of things to worry about is a bit longer than you thought, or whether it indicates that we will not typically succeed in providing effective protection at this time.
raman: Out of band info being put into local storage, will soon be put into cloud storage. I predict all local storage will move to cloud
... Users have conflicting requirements
... Bad guys will use same hooks to do bad things
<noah> NM: I tend to believe the latter. That being the case, we need to then decide whether it's worth asking people to plug at least some of the holes. My leaning is "yes, sometimes", Henry seems to say "probably not", but I think that's the next discussion we should have.
raman: This is a wakeup call. No easy answer
dka: Looks like a false dichotomy. Either you admit web privacy is dead, or you plug a million holes
... There could be tactical approaches. Plug some holes, help a bit, help users to make informed decisions
... also longer term work to be done on privacy, much bigger than TAG, but TAG can play a role, workshops/intitiatives
<noah> . ACTION: Appelquist to prepare early draft of TAG thoughts on implications of Evercookie. Due: 2010-12-08
dka: We can show leadership by saying something [appropriate / reasoned / limited] about evercookie
<noah> ACTION: Appelquist to prepare early draft of TAG thoughts on implications of Evercookie. Due: 2010-12-08 [recorded in http://www.w3.org/2001/tag/2010/10/21-minutes#action02]
<trackbot> Created ACTION-489 - Prepare early draft of TAG thoughts on implications of Evercookie. Due: 2010-12-08 [on Daniel Appelquist - due 2010-10-28].
yves: It's more like spam. Never ending battle. But we have to keep fighting.
<noah> If my privacy is protected as well as I am insulated from spam, then my life is one big open book.
yves: Explaining issue is important
... it's a tax
masinter: 'Privacy' is not the quantity we're trying to optimize
... Goal is to match what's happening with what's expected
... I send email to a list that I think is limited readership, is it really?
... Put evercookies in that context and I'll be happy
noah: (administrative interrupt as DKA departs) Considering skipping call next week (TPAC is the week after that)
ashok: Encrypt everything at a low level?
jar: Hal [Abelson] has been looking at tagging-based hardware to support accountability
noah: Unknown people can't be held accountable
... they're piecing together information about me
masinter: We haven't done the threat analysis. Don't know what problem to solve
noah: ...
masinter: They have my IP address and HTTP headers, what's new?
ht: The evercookies point was that the vulnerabilities were already there
yves: Companies are buying one another, and otherwise sharing information with one another
... so you get correlations
masinter: Clearing your cookies not only doesn't get you what you need, it gets in the way of getting what you want
<Zakim> ht, you wanted to make the strategy point
(people use 'clear cookies' to logout / clear credentials, and it often works)
ht: As far as new activity in this area might go, it shouldn't confuse privacy with protecting private data
... If there are thought leaders giving a message that's the right one, then we (W3C) could get behind it
noah: We have enough actions, yes?
<noah> ACTION: Noah and others(?) going to privacy workshop to report back to the TAG? [recorded in http://www.w3.org/2001/tag/2010/10/21-minutes#action03]
<trackbot> Created ACTION-490 - And others(?) going to privacy workshop to report back to the TAG? [on Noah Mendelsohn - due 2010-10-28].
jar: Given that we're not on top of thinking on this, how about if someone gets educated by reading something (e.g. about Hal's ideas)
<noah> ACTION-490 Due 2010-12-21
<trackbot> ACTION-490 And others(?) going to privacy workshop to report back to the TAG? due date now 2010-12-21
A#B where A -> C#D
yves: Taxonomy of fragids
... FIrst case: book chapter. Also SVG file with multiple icons, point to a particular icon
... I tested the SVG fragment case and it worked (and it's in the spec)
timbl: In the python API, you don't have enough information
... x = urlib.urlopen('foo')
... s = x.read()
... x.headers
jar: I think Julian was saying the whole discussion is moot, since the decision was made 10 years ago and published in the errata.
... So the time to review it would have been then. Cat's out of the bag.
<Yves> http://skrb.org/ietf/http_errata.html
jar: (To Yves) Then this SVG case is the example that I had asked for?
(everyone trying to track down ietf commitment to this erratum)
(looking at archive's version of skrb.org)
<Yves> http://lists.w3.org/Archives/Public/ietf-http-wg-old/1999MayAug/0115.html
ht: Possible courses of ACTION: (1) Revert this change to Location:. Does harm, no evidence of use. (2) We don't know if anyone's using it, we don't have fragment combination rules, leave it in. (3) Here are frag combination rules that cover particular cases, don't do it otherwise (i.e. negative consequences will follow).
masinter: (4) You can have one fragment id, but not two.
ht: The people who set up a redirect are not the same as the people who capture the fragid URI
... Or, people publish docs with names in them. I select something that seems to have a name, click view fragment source, put together a URI, and send it on to someone.
... It's not necessarily the case that there is coordination
masinter: You can do this, but something might break.
masinter (reworded by jar): If you deploy a 30x Location: C#D, then be aware that anyone who creates a URI A#B, might be inconvenienced (since there are no fragment combination rules).
scribe: (that is, A 30x redirects to C#D)
<timbl> http://purl.org/ontology/bibo/Book
masinter: There's a tendency to want to control both sides of the conversation - MUSTs that apply to parties not constrained by the spec in question
<timbl> $ curl -I http://purl.org/ontology/bibo/Book
<timbl> HTTP/1.1 302 Moved Temporarily
<timbl> Location: http://bibotools.googlecode.com/svn/bibo-ontology/tags/1.3/bibo.xml.owl#Book
yves: Case 2. Absolute fragment - exactly one place in the document.
... Compare xpointer, from one point in a doc, select something just following, not same as selecting from whole document.
<timbl> >>> import urllib
<timbl> >>> x = urllib.urlopen("http://purl.org/ontology/bibo/Book")
<timbl> >>> s = x.read()
<timbl> >>> s
<timbl> '<!DOCTYPE HTML PUBLIC "-//IETF//DTD HTML 2.0//EN">\n<html><head>\n<title>400 Bad Request</title>\n</head><body>\n<h1>Bad Request</h1>\n<p>The request was invalid: the URI included an un-escaped hash character</p>\n</body></html>\n'
<timbl> >>>
yves: Example with RDF: fragment defines something that's in the document. Or javascript, where it's a script state.
... It's not really in the realm of the mime type
raman: The way I tried to reconcile this, think of it as an argument to a client side function that deals with the fragid. In general, it's an arg to some function on the page. If no such, then look for element.
... Server/client not completely separate, since client is running code that came from server
yves: Unpredictability leads to impossibility of a universal way to define fragid combination.
<Yves> http://lists.w3.org/Archives/Public/www-tag/2010Oct/0036.html
yves: Best solution is to drop the original one. That's what most browsers do, anyhow.
masinter: What if you redirect to a PDF, that has its own idea of fragids.
ht: Right, that's the reason to drop the first fragid
yves: In case of redirect A#B to C, to be consistent you'd have to drop the #B
noah: When I go to W3C, and I ask you from foo (instead of foo.html), and that redirected to foo.html ... lots of places where keeping the #B fragment is leveraged
ht: Maybe I was too quick to agree with Yves on the media type question. The naive view is that the document is in a different place. But really it's more like conneg
... It's still reasonable to apply the original fragid (#B in A#B when A redirects to C#D)
<ht> In either case the presumption of consistent fragid interpretation is legitimate
masinter: Don't redirect to [already typed into irc]
<ht> This justifies applying 'B' to C when A#B redirects to C
masinter: If you do conneg, don't do it where fragids mean different things
<Yves> http://www.w3.org/TR/2004/REC-webarch-20041215/#media-type-fragid
<Zakim> noah, you wanted to talk about see also
noah: In 303, second doc is clearly different, but 301 says the docs are pretty much the same
ht: 90% cases. Where no fragid combination, we agree with both situations, A->C#D and A#B where A->C.
... In the fragid combination case, it's allowed... but health warning.
<noah> Therefore for A#B --303 Redirect to--> C#D, applying #B to either C or (somehow) to C#D is highly suspect.
scribe apologizes, missing this
didn't scribe conversation between henry and tim
ht: propose: The TAG will not request HTTPbis WG to restrict Location: to absolute URIs
... OK, that's a starting point
<Yves> (for the scribe it is http://trac.tools.ietf.org/wg/httpbis/trac/ticket/6 )
timbl: Agree only on condition... that there is a sufficient health warning
yves: Notes that Larry proposed such a health warning
ht: Let's table this
<timbl> ____ [begin demo] ____
<timbl> $ curl -L http://purl.org/ontology/bibo/Book
<timbl> <!DOCTYPE HTML PUBLIC "-//IETF//DTD HTML 2.0//EN">
<timbl> <html><head>
<timbl> <title>400 Bad Request</title>
<timbl> </head><body>
<timbl> <h1>Bad Request</h1>
<timbl> <p>The request was invalid: the URI included an un-escaped hash character</p>
<timbl> </body></html>
<timbl> $
<timbl> ____ [end demo] ____
jar: not willing to assent to Larry's warning today. May later decide it's ok
noah: NOTE: Larry has made a proposal. We will take this up on a call.
<ht> ____ [begin demo] ____
<ht> wget http://purl.org/ontology/bibo/Book
<ht> --2010-10-21 15:17:26-- http://purl.org/ontology/bibo/Book
<ht> Resolving purl.org (purl.org)... 132.174.1.35
<ht> Connecting to purl.org (purl.org)|132.174.1.35|:80... connected.
<ht> HTTP request sent, awaiting response... 302 Moved Temporarily
<ht> Location: http://bibotools.googlecode.com/svn/bibo-ontology/tags/1.3/bibo.xml.owl#Book [following]
<ht> --2010-10-21 15:17:26-- http://bibotools.googlecode.com/svn/bibo-ontology/tags/1.3/bibo.xml.owl
<ht> Resolving bibotools.googlecode.com (bibotools.googlecode.com)... 209.85.227.82
<ht> Connecting to bibotools.googlecode.com (bibotools.googlecode.com)|209.85.227.82|:80... connected.
<ht> HTTP request sent, awaiting response... 200 OK
<ht> ____ [end demo] ____
<noah> close ACTION-462
<trackbot> ACTION-462 Write draft of best practices on redirection for secondary resources (with help from Tim) closed
masinter: Opposed to telcon time on this
<noah> ACTION: Noah to schedule telcon attempt to formulate health warning on secondary resource redirection noting Larry proposal in 21 Oct 2010 F2F record [recorded in http://www.w3.org/2001/tag/2010/10/21-minutes#action04]
<trackbot> Created ACTION-491 - Schedule telcon attempt to formulate health warning on secondary resource redirection noting Larry proposal in 21 Oct 2010 F2F record [on Noah Mendelsohn - due 2010-10-28].
<johnk> break
ACTION jar to Review Larry's health warning on redirection to secondary resources and either agree or fix
<trackbot> Created ACTION-492 - Review Larry's health warning on redirection to secondary resources and either agree or fix [on Jonathan Rees - due 2010-10-28].
http://www.w3.org/2001/tag/group/track/actions/pendingreview
<noah> CLOSE ACTION-473
<trackbot> ACTION-473 Draft possible TAG response on HTML extensibility closed
johnk: I deferred announcement of my writing on action-417 pending further work on web apps
ht: If I were at TPAC I'd like to drill on the DOM-based objections to 'just like SVG'
<noah> http://www.w3.org/2001/tag/group/track/actions/overdue?sort=owner
Take action-473 to email
action-454?
<trackbot> ACTION-454 -- Daniel Appelquist to take lead in organizing outside contacts for TAG F2F -- due 2010-10-05 -- OPEN
<trackbot> http://www.w3.org/2001/tag/group/track/actions/454
<noah> close ACTION-454
<trackbot> ACTION-454 Take lead in organizing outside contacts for TAG F2F closed
<noah> ACTION-116 Due 2011-02-11
<trackbot> ACTION-116 Align the tabulator internal vocabulary with the vocabulary in the rules http://esw.w3.org/topic/AwwswDboothsRules, getting changes to either as needed. due date now 2011-02-11
action-280?
<trackbot> ACTION-280 -- John Kemp to (with John K) to enumerate some CSRF scenarios discussed in Jun in Cambridge -- due 2010-10-11 -- OPEN
<trackbot> http://www.w3.org/2001/tag/group/track/actions/280
johnk: Had trouble figuring out which scenarios were meant.
jar: Sounds like a minutes failure
johnk: Couldn't find the 2x2 table that Dan wrote on the board
... I'll add a note to section 7
<johnk> I'll add a note to ACTION-417 noting that it should include describing CSRF scenarios
<noah> Closing ACTION-280 because will pursue CSRF under ACTION-416
<noah> close ACTION-280
<trackbot> ACTION-280 (with John K) to enumerate some CSRF scenarios discussed in Jun in Cambridge closed
ashok: Has there been any progress on CORS/UMP?
jar: many months ago. new WG in the works maybe
<noah> ACTION-355?
<trackbot> ACTION-355 -- John Kemp to explore the degree to which AWWW and associated findings tell the interaction story for Web Applications -- due 2010-10-05 -- OPEN
<trackbot> http://www.w3.org/2001/tag/group/track/actions/355
<noah> JK: Made some progress on that
johnk: Cases not covered by AWWW - some covered in email
... but it's not far enough along for me to want to review
<noah> ACTION: Noah to schedule discussion of interim work on ACTION-355 Due: 2010-11-09 [recorded in http://www.w3.org/2001/tag/2010/10/21-minutes#action05]
<trackbot> Created ACTION-493 - Schedule discussion of interim work on ACTION-355 Due: 2010-11-09 [on Noah Mendelsohn - due 2010-10-28].
<noah> ACTION-416?
<trackbot> ACTION-416 -- John Kemp to work on diagrams in "From Server-side to client-side" section of webapps material -- due 2010-10-11 -- OPEN
<trackbot> http://www.w3.org/2001/tag/group/track/actions/416
<johnk> http://www.w3.org/2001/tag/2010/10/interaction-examples.html
johnk: I want to know what we're doing with webapps
noah: We decided to dive deep on storage, state, and apis
ashok: and we need a short overview based on stuff larry sent
masinter: We've talked about a framework / broad outline; and a few things that fit into it
<noah> also diving on privacy
noah: Where to draw the line between work on storage and work on privacy?
<DKA> IMO privacy considerations should be part of storage document.
masinter: I'd like a way to think about relationship between client side and server side processing and how things move from one to the other
johnk: I put some diagrams up, see http://www.w3.org/2001/tag/2010/04/WebApps.html
masinter: diagrams are fine for now. let's move on
ashok: I thought the diagrams with words around them would constitute an intro to the web apps work.
timbl: Semantics of diagrams unclear
<noah> ACTION-416 Due 2011-03-01
<trackbot> ACTION-416 Work on diagrams in "From Server-side to client-side" section of webapps material due date now 2011-03-01
<noah> Pushing out 416 pending better idea of what if anything is needed.
<noah> ACTION-355 Due 2011-01-02
<trackbot> ACTION-355 Explore the degree to which AWWW and associated findings tell the interaction story for Web Applications due date now 2011-01-02
<noah> ACTION-472?
<trackbot> ACTION-472 -- Larry Masinter to update the mime-draft, due 2010-10-12 -- due 2010-10-07 -- OPEN
<trackbot> http://www.w3.org/2001/tag/group/track/actions/472
<noah> ACTION-472 Due 2011-12-01
<trackbot> ACTION-472 Update the mime-draft, due 2010-10-12 due date now 2011-12-01
<noah> ACTION-442?
<trackbot> ACTION-442 -- Ashok Malhotra to try and integrate NM and TVRs words into a coherent draft -- due 2010-09-09 -- OPEN
<trackbot> http://www.w3.org/2001/tag/group/track/actions/442
<noah> AM: Overtaking by ACTION-481 or 482
<noah> close ACTION-442
<trackbot> ACTION-442 Try and integrate NM and TVRs words into a coherent draft closed
<noah> ACTION-468?
<trackbot> ACTION-468 -- Noah Mendelsohn to invite Thinh to telcon where Tim will be available to discuss "forbidding of hyperlinking" -- due 2010-10-05 -- OPEN
<trackbot> http://www.w3.org/2001/tag/group/track/actions/468
<noah> Tim is likely 3 Dec and 10 Dec
<noah> ACTION-468?
<trackbot> ACTION-468 -- Noah Mendelsohn to invite Thinh to telcon where Tim will be available to discuss "forbidding of hyperlinking" -- due 2010-10-05 -- OPEN
<trackbot> http://www.w3.org/2001/tag/group/track/actions/468
<noah> Tim OK 18 Nov, 2 Dec
<noah> ACTION-468 Due 2010-11-16
<trackbot> ACTION-468 Invite Thinh to telcon where Tim will be available to discuss "forbidding of hyperlinking" due date now 2010-11-16
<noah> close ACTION-474
<trackbot> ACTION-474 Ask Raman to set up phone at F2F closed
<noah> close ACTION-465
<trackbot> ACTION-465 Schedule F2F discussion of ACTION-282, "which metadata mechanisms to use when". Get reading list from Larry and www-tag. closed
<noah> ACTION-429
<noah> ACTION-201
<noah> ACTION-201?
<trackbot> ACTION-201 -- Jonathan Rees to report on status of AWWSW discussions -- due 2010-10-05 -- OPEN
<trackbot> http://www.w3.org/2001/tag/group/track/actions/201
<noah> JAR: Will do for next F2F
action-201 due 2011-01-25
<trackbot> ACTION-201 Report on status of AWWSW discussions due date now 2011-01-25
<noah> close ACTION-429
<trackbot> ACTION-429 Work on arrangements for a TAG F2F, probably @ Google early Oct (self-assigned) closed
<noah> Noticed that ACTION-482 is redundant with ACTION-475
<noah> close ACTION-482
<trackbot> ACTION-482 Write a draft on client-side storage with help from DanA Due: 2010-11-30 closed
<noah> close ACTION-485
<trackbot> ACTION-485 Have Amy reserve rooms for TAG F2F Feb 8-10 2011 closed
<noah> close ACTION-483
<trackbot> ACTION-483 Inform TAG of TPAC meeting schedule closed
<noah> close ACTION-486?
<noah> close ACTION-486
<trackbot> ACTION-486 Schedule group discussion of goals and roadmap for metdata work closed
<noah> ACTION-381?
<trackbot> ACTION-381 -- Jonathan Rees to spend 2 hours helping Ian with http://www.w3.org/standards/webarch/ -- due 2010-11-01 -- OPEN
<trackbot> http://www.w3.org/2001/tag/group/track/actions/381
<noah> ACTION-381 Due 2010-12-01
<trackbot> ACTION-381 Spend 2 hours helping Ian with http://www.w3.org/standards/webarch/ due date now 2010-12-01
<noah> ACTION-487?
<trackbot> ACTION-487 -- Jonathan Rees to assess potential impact of IRI draft on RDF/XML, OWL, and Turtle -- due 2011-12-01 -- OPEN
<trackbot> http://www.w3.org/2001/tag/group/track/actions/487
<noah> ACTION-478?
<trackbot> ACTION-478 -- Jonathan Rees to prepare a first draft of a finding on persistence of references, to be based on decision tree from Oct. F2F Due: 2010-01-31 -- due 2010-10-27 -- OPEN
<trackbot> http://www.w3.org/2001/tag/group/track/actions/478
<noah> ACTION-476?
<trackbot> ACTION-476 -- Jonathan Rees to draft a short note to 3023bis editors reflecting the discussion / consensus... -- due 2010-10-26 -- OPEN
<trackbot> http://www.w3.org/2001/tag/group/track/actions/476
<noah> ACTION-23?
<trackbot> ACTION-23 -- Henry S. Thompson to track progress of #int bug 1974 in the XML Schema namespace document in the XML Schema WG -- due 2010-11-30 -- OPEN
<trackbot> http://www.w3.org/2001/tag/group/track/actions/23
<timbl> I wonder whether http://www.w3.org/2001/tag/group/track/issues/46 has been overtaken by events - no one is doing XML 1.1, only XML 1.0 erratum
ADJOURNED
