[Standards] XEP-0277: no link to original post in comment item + published/updated

Sergey Dobrov binary at jrudevels.org
Wed May 6 07:08:38 UTC 2015

On 05/05/2015 23:01, Goffi wrote:
> On 05/05/2015 16:59, Sergey Dobrov wrote:
>> On 01/05/2015 21:28, Goffi wrote:
>> Sorry I was not replying, I'm quite busy now, but am going to join you
>> guys ASAP, unfortunately, got a flu now.
> No worries, nice to see you back.
>> Yes, I am aware of the issue but it can't be fixed before we have a good
>> way to store XML in the metadata.
> OK
actually, when we got such a place some links like a comment node will 
go there too.

>>> 2) published/updated
>>> the <published/> and <updated/> elements are given by the publisher, he
>>> can write anything he wants inside. I think that the server should
>>> enforce these data, or we will have the same issues as in email like
>>> mesages in the future (e.g. to appear first).
>> Yes, I know about this. This was made purposely to make it possible, for
>> instance, to import old blog posts from different source.
> Would it be possible to allow enforcement as an option ? An admin can
> decide to import old blogs, but an user should not be able to fake it,
> or at least not in every situations.
I don't see a necessity and possibility to do that: it will require 
pubsub be aware of payloads format which is not suitable, I guess.

If you think it's necessary on your service you can force it on your 
pubsub but I don't see why to add such a restriction on the XEP level, 
it's just too strict (and not really possible for 277).

> I think an implementation note should say that the publisher or updated
> field can be modified by the pubsub service, depending on the
> implementation/configuration. This way a client can expect to have a
> published/updated field changed by the server.
Well, we need another filtering entity here (I said about necessity of 
it long time ago), which also will be used to filtering spam (or obscene 
words), this way client should always expect that the payload can be 
modified, but again this is not about 277, I believe.

> cheers
> Goffi

More information about the Standards mailing list