19 May 2021


Cristiano_Aguzzi, Daniel_Peintner, Ege_Korkan, Kaz_Ashimura, Klaus_Hartke, Michael_Koster, Michael_McCool, Taki_Kamiya, Tomoaki_Mizushima
kaz, taki, TK

Meeting minutes


<kaz> May-12

Taki: minutes were accepted from last week.

Debriefing from the main call

<kaz> TD discussion during the main call

McCool: abstract intro and change log were merged.

McCool: made correction to abstract.

McCool: changed wording mistake.

McCool: basically ready waiting for sebastian's change to changelog.

PR# 1147

<kaz> PR 1147 - PR including the comments

Kaz: there are only two lines mentioning "features at-risk" but is that OK?

McCool: At risk part needs to be taken care of before CR.

Kaz: this is for an updated WD not CR, so should be OK

Taki: PR# 1147 was merged.

Kaz: my only question is relationship between this PR's update and the candidate draft

candidate draft for updated WD

Kaz to check if the spec is clean on wd-update-candidate2 branch.

Daniel: I still see broken links in the spec.

Kaz: we can check in static version first, then fix broken links.

<kaz> Link checker

Kaz: DP can fix links directly on the branch.

Daniel: Should it be which branch?

Kaz: ideally to both branches, but you can create yet another branch for publication like Taki did for "wd-update-candidate2"

Daniel: I will ping others after fixing links.

McCool: Fix rendered version, and merge the changes back in.

Kaz: Please take a look at possible errors, and can discuss what to do else on Friday.

Use case document clean up was already done by Kaz, but PR to be done for applying the changes back to the respec draft.

Issue 1044

<kaz> Issue 1044 - Adding term to indicate a stream of data

McCool: Chunked transfer is needed in directory use case.

McCool: HTTP use case needs to be able to handle it.

Ege: devices that generates big data can choose to output as a stream.

McCool: HTTP 1.1 and 2.0 have different mechanisms.

McCool: images and videos are big, and videos are unlimited length.

Koster: chunking can be used on HTTP, but chunking itself does not mean streaming.

McCool: pagination and chunking are different.

McCool: chunking does not mean structure.

Cristiano: HTTP provides chunking. if protocol does not support chunking, then server still want to make chunks, so streaming can be useful.

McCool: "streaming" is a hint.

McCool: we can check contenttype can use any options for chunking.

McCool: such as streaming=true

Koster: it depends on protocols. HTTP has a way to do.

McCool: HTTP2 uses different mechanism.

McCool: we can capture this dicussion in the issue.

McCool: we can also take a look at alternatives such as contenttype.

Koster: CoAP only has blocking.

Koster: but it is not continuous.

Cristiano: streaming can make TD handling complex.

McCool: protocol binding can also cover this aspect.

McCool: each protocol can define what streaming=true means.

Ege: When I opened the issue, I heard some people need this. We can keep this issue open and try to solve it going forward.

Kaz: we can clarify media use cases first (like video streaming using MEPG-DASH or HLS) but that would be a complex task. We can address this in 2.0.

Cristiano: scripting support is just basic now. we can wait till 2.0.

<kaz> [adjourned]

Minutes manually created (not a transcript), formatted by scribe.perl version 131 (Sat Apr 24 15:23:43 2021 UTC).