Mobile Web Best Practices Working Group Teleconference

One time URIs

<francois> Heiko's email

Francois: Raised by Heiko
... We cannot address it more than we already do.

Heiko: For one time URLs we can add no-transform.

<francois> PROPOSED RESOLUTION: One-time URIs are already addressed in the guidelines. Close the discussion.

<hgerlach> +1

<francois> +1

<rob> +1

RESOLUTION: One-time URIs are already addressed in the guidelines. Close the discussion.

Idempotency

<francois> fd's email

Francois: We had the correct definition of idempotency.
... If you do the same action multiple times you get the same result.
... GET requests are supposed to be idempotent.

<francois> PROPOSED RESOLUTION: Keep things as resolved re. idempotency for the end of 4.1.2. Valid use of "idempotent".

<rob> +1

+1

<francois> +1

RESOLUTION: Keep things as resolved re. idempotency for the end of 4.1.2. Valid use of "idempotent".

<hgerlach> +1

<francois> Close ACTION-751

<trackbot-ng> ACTION-751 Make sure that we are clear about idempotency and side-effect freedome etc. per Dom's original illuminating point about this section closed

Sessions and consistency of user experience

Francois: One things we do not say a lot about is URL-rewriting
... a click on a link goes through the CT proxy again.
... Don't think we need to say much about this.

Heiko: If you do not have CT enabled, you can't insert links.
... Should we define an additional header?
... CT proxy whould add the header.

<hgerlach> +q

Francois: When a page is not transformed, then user cannot be back to CT proxy easily
... also, we do not say much in the guidelines about rewritten URLs and that the CT proxy needs to set the HTTP headers like the request actually came from the client.
... don't think we need to emphasize this in the document.

Heiko: CT proxy can be set up as an HTTP proxy where URLs are not rewritten.

Francois: Don't think we can say much about CT proxy and directing to a page not going through the CT proxy.
... I see headers and footers as out of scope.
... Cache-control needs to be strong...could use out of scope means to rewrite URLs.

Heiko: We have two different no-transform options.
... first can be HTTP header, second can be in META tag.

<hgerlach> http://wapreview.com/blog/?p=516

<francois> ACTION: daoust to check on equivalence between an HTTP header and a META element in HTML pages [recorded in http://www.w3.org/2008/05/27-bpwg-minutes.html#action01]

<trackbot-ng> Created ACTION-765 - Check on equivalence between an HTTP header and a META element in HTML pages [on François Daoust - due 2008-06-03].

Francois: I don't think we can do anything about this issue.
... there is an an editorial note in the document.

Rob: Maybe the note should stay there.

Francois: Maybe we could put in a note about how the user experience could break with no-transform.

Rob: Have a note that it would be a good idea to keep the user experience the same across the session.

<hgerlach> how do you define session#

Francois: Don't think we should use "session" because then we would need to define session.

<francois> Note: When possible, the consistency of user experience should be maintained across a sequence of related requests.

<francois> PROPOSED RESOLUTION: for sessions: replace the editorial note with the above note. Do not mention sessions. Do not go into more details.

<rob> +1

<hgerlach> +1

<francois> 0

RESOLUTION: for sessions: replace the editorial note with the above note. Do not mention sessions. Do not go into more details.

<francois> Close ACTION-755

<trackbot-ng> ACTION-755 Summarize the issue that may be raised by sessions closed

Link element

<francois> discussion

Francois: Has been discussed multiple times.
... Aaron is going to say more about this when he has time.
... LINK element links to alternate representions of a resource.
... HTML page could link to a handheld version of the page.
... Problem that is raised is when the handheld version of the page is requested, there is no way to tell if the page is handheld.
... There is a way to use the LINK element to "kind of link to itself" but I'm not sure this is used in practice.

<hgerlach> -q

Francois: The LINK header come back to life sometime. Could be used for non HTML pages.
... Related to long running issue 222.
... Jo wanted to make a comment on issue 222.

Sean: Treat LINK element as a redirect to mobile version.

Rob: Tend to agree.

Francois: Wait until next week to make a resolution on this.

Cache-Control: no-transform

<francois> Discussion

Francois: The no-tranform switch needs to be respected.
... Most done on the editorial notes.
... A few more things we need to agree on.
... Want to do last call around F2F date.
... If there is anything else in the document we want to change, now is the time.
... once it goes to last call, working group should agree and comments should come from the outside.