Talk:Microsummaries
Note: I would prefer for discussion about Microsummaries to happen in the mozilla.dev.apps.firefox newsgroup, as I find this forum non-conducive to followable threaded discussion. -myk
Dumb idea?
Are these microsummaries going to generate a page request every time a user opens their browser, or even more often? That's going to generate one page request for every bookmarked page per user. That's a hell of a lot of page requests globally, potentially costing anyone using this technology a LOT in terms of bandwidth. Unless the interface to this actually adds some big benefit to site publishers in the form of getting more peoples' eyes on their sites, this is going to be a wasted effort.
What's wrong with RSS to notify people of changes? Why are you trying to add RSS type functionality into HTML and into the Bookmarks system? Is this actually needed?
Processing Instruction
It would be nice if an XML Processing Instruction was added as a way of associating a micro-summary with an XML resource. Furthermore, a content type for the format should be established.
<?micro-summary type="application/micro-summary+xml" href="/summary.xml"?>
The best part of using a processing instruction is that it works in all XML languages without breaking anything if the client application doesn't support the feature. Daniel Schierbeck 05:23, 20 Mar 2006 (PST)
RDF
What about using any published RDF (and not just RSS) as an input for microsummaries? For that matter, generic XML should also be supported as an input format, e.g. the MMOG EVE Online publishes character information as XML.
~Grauw
Update period
As suggested by stonedyak on MozillaZine forums, the spec should provide a way to dictate how often a microsummary is updated.
The interval could be dictated in seconds, minutes or whatever seems appropriate. This would give added flexibility for generator authors, but would make the microsummary service more efficient due to running fewer useless checks.
Example:
<?xml version="1.0" encoding="UTF-8"?> <generator xmlns="http://www.mozilla.org/microsummaries/0.1" name="Yahoo! Finance Stock Quote"> <update interval="14400"/> <pages> <include/> </pages> <template> [...] </template> </generator>
--Cusser 12:38, 14 May 2006 (PDT)
As a site developer, I definitely would want some reassurance that I'm not going to be slammed with a constant stream of requests. So while I don't know if I need the ability to customize it, I need some kind of information. Does this happen constantly while the user's browser is open? How often? Does it happen every time the bookmarks menu is opened? Is there any kind of bound? What happens in failure - does it retry? what's the backoff mechanism?
--Slamb 16:40, 17 October 2006 (PDT)
Privacy
As a user, I'd like to know - what information does this feature leak to site developers? Do the requests happen the whole time my browser is open? Do they contain my cookies at the site? Actually, even if the answer to the last question is "no", the URL could be unique to me, web bug-style. Is there a way to turn the feature off? An easy way to know if a site is using it?
--Slamb 16:40, 17 October 2006 (PDT)