[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
RE: [syndication] Re: BBC News RSS
Whoops, nearly missed this...
> Danny Ayers wrote:
> > Personally I'd advise using a different format altogether. Any.
>
> So is Ideagraph going to not import OPML, making it useless to anyone who
> has ever used an aggregator before? And will it not export OPML, making it
> the sort of dead end that anyone with a bit of sense and experience with
> aggregators will avoid?
So anything that doesn't use OPML isn't any use to people that have used
aggregators before? Dave has achieved his format lock in? Heh, *not*. Mostly
because existing tools don't support OPML consistently - my last experience
was export from RSS Bandit, import to Bottom Feeder - result: zilch. Sense
and experience tells me that anything's better than OPML, hence my advice.
I have played with OPML import, using SAX, XSLT and crude regexp-style
parsing, and that is why I wouldn't recommend it. I'll probably support a
narrow subset, with a minimal bit of detection code to record that the
imported data was OPML, and hence could include unreadable stuff. If I have
time then similarly I'll probably put together a wrapper on export to cover
a subset of OPML. The data might be useful. Maybe. But don't expect me to be
happy about it or wave the flag. I'm not going to jump through hoops to
support endless variations on an unspecified theme.
> Battle cries in the Syndication Wars are fine when you know that everyone
> listening knows that that's all they are, but in the other world,
> beyond the
> battlefield, virtually every aggregator around will import and
> export OPML,
> in one flavor or another, to one extent or another. OCS is much prettier,
> and virtually unsupported.
The cries aren't battle cries, just cries of anguish following time spent
with OPML. It hurts to have to support this mess, when there are clear
alternatives.
There may not be many newsreader/aggregators that support OCS. But every
tool that uses RDF has support for it. If we allow the agenda to be narrowed
to what's easy using formats like OPML then we're in a serious rut.
In any case I don't really think the fact that all today's aggregators
support OPML carries as much weight as you imply - they're pretty much all
stuck at the same level of sophistication, due in large part to the fact
that it's not straightforward to extend such formats. Have we really reached
the limit of what can be done? Glorified electronic newpapers? Gimmee a
break.
> Valuable contribution to the syndication community: search out as many
> examples as you can find, detail the differences, put it on a wiki.
What's the point? Any attributes are permitted, so it's a task that's
potentially open ended. I'd rather spend my time developing the horseless
carriage than following carthorses with a bucket. *
> One more straw tilting me toward unsubscribing from my very last
> syndication-related email list: well, you know ;)
I hadn't checked this list for a week or so, so I guess I must be feeling
the same way. Maybe I'll unsubscribe for a year or two, come back when the
evolutionary fork has reached extinction.
Cheers,
Danny.
* unfair on horses, sorry equine dudes