RSS as a pseudo Web Service

[bgranam]

Greetings.

I've been musing about the possibility of using RSS as a kind of
pseudo Web Service and would like some opinions on the concept.

RSS would be used as a transport layer for arbitrary RDF content
between two or more computers. The RDF content would not be intended
for direct display, but would be parsed out with XSLT and fed into
another program for further processing.

Good, bad, stupid, brilliant? Your thoughts would be appreciated.

Kind regards,
Brian Grainger








Yahoo! Groups Links

<*> To visit your group on the web, go to:
http://groups.yahoo.com/group/rss-dev/

<*> To unsubscribe from this group, send an email to:
rss-dev-unsubscribe-***@public.gmane.org

<*> Your use of Yahoo! Groups is subject to:
http://docs.yahoo.com/info/terms/

[James Holderness]

I'd recommend against it. You're assumedly going to have a server at one end
that's not going to be an RSS server, a client at the other end that's not
going to be an RSS client, so why use RSS for the protocol between the two?

If you really think the RSS protocol fits your needs perfectly and you're
not just stuffing your data into RSS because you don't want to come up with
a format of your own, at least give it a different top level element so you
don't run the risk of someone coming across your feed and trying to
subscribe to it.

IMHO.

Regards
James

Post by bgranam
I've been musing about the possibility of using RSS as a kind of
pseudo Web Service and would like some opinions on the concept.
RSS would be used as a transport layer for arbitrary RDF content
between two or more computers. The RDF content would not be intended
for direct display, but would be parsed out with XSLT and fed into
another program for further processing.
Good, bad, stupid, brilliant? Your thoughts would be appreciated.

Yahoo! Groups Links

<*> To visit your group on the web, go to:
http://groups.yahoo.com/group/rss-dev/

<*> To unsubscribe from this group, send an email to:
rss-dev-unsubscribe-***@public.gmane.org

<*> Your use of Yahoo! Groups is subject to:
http://docs.yahoo.com/info/terms/

[Bill Kearney]

Indeed, why bother wrapping it in RSS at all? If all you're passing is RDF
data then why not simply pass it using REST principles? Just pull the data
via HTTP.

But let's step back, is there some other parallel reason for wanting to use
RSS? When using RSS-1.0 it's entirely possible to 'carry along' all manner
of other data. Only an RDF-aware application would bother paying attention
to it. It'd be 'perfectly safe' to do it from the perspective of not
disrupting RSS client behaviors. One downside would be wasted bandwidth.
Why bother cramming extra data into the feed when no normal feed processing
tool would consume it? I'm not talking basic metadata like timestamps or
category data as more of that would certainly help. But other stuff pulled
by a ton of RSS clients would waste the bandwidth on both ends.

If it's private data (not intended for public subscriptions) and you've got
reasonable controls over the client-side users then bandwidth probably isn't
much of an issue, other than waste-not/want-not principles of course.

As to processing RDF via XSLT you're in for a whole other set of
adventures...

-Bill Kearney
Syndic8.com

Post by James Holderness
I'd recommend against it. You're assumedly going to have a server at one end
that's not going to be an RSS server, a client at the other end that's not
going to be an RSS client, so why use RSS for the protocol between the two?
If you really think the RSS protocol fits your needs perfectly and you're
not just stuffing your data into RSS because you don't want to come up with
a format of your own, at least give it a different top level element so you
don't run the risk of someone coming across your feed and trying to
subscribe to it.

Post by bgranam
I've been musing about the possibility of using RSS as a kind of
pseudo Web Service and would like some opinions on the concept.
RSS would be used as a transport layer for arbitrary RDF content
between two or more computers. The RDF content would not be intended
for direct display, but would be parsed out with XSLT and fed into
another program for further processing.
Good, bad, stupid, brilliant? Your thoughts would be appreciated.

Yahoo! Groups Links

<*> To visit your group on the web, go to:
http://groups.yahoo.com/group/rss-dev/

<*> To unsubscribe from this group, send an email to:
rss-dev-unsubscribe-***@public.gmane.org

<*> Your use of Yahoo! Groups is subject to:
http://docs.yahoo.com/info/terms/

[g***@public.gmane.org]

Post by Bill Kearney
Indeed, why bother wrapping it in RSS at all? If all you're passing is
RDF data then why not simply pass it using REST principles? Just pull
the data via HTTP.

A good question, which is why I wanted to enquire. You've now made me aware
of REST, which I wasn't before, so I will certainly take a closer look at
it.

Post by Bill Kearney
But let's step back, is there some other parallel reason for wanting to
use RSS?

None, other than it would be RDF, like the content.

Post by Bill Kearney
As to processing RDF via XSLT you're in for a whole other set of
adventures...

Yes, that is so.

Regards,
Brian
--
Using Opera's revolutionary e-mail client: http://www.opera.com/mail/






Yahoo! Groups Links

<*> To visit your group on the web, go to:
http://groups.yahoo.com/group/rss-dev/

<*> To unsubscribe from this group, send an email to:
rss-dev-unsubscribe-***@public.gmane.org

<*> Your use of Yahoo! Groups is subject to:
http://docs.yahoo.com/info/terms/

[if]

Although I take the point which about RSS not being ideal for data
interchange. There are some good reasons to use RSS over other methods.

1. Its pretty easy to parse and there are many tools to help
2. You can serve 2 or more audiences, so our RSS feed is quite heavy and
contains information which people can read using a RSS reader or parse
and use in there own applications.
3. Its very common, so many clients now want RSS over anything else (at
least thats true in the news market)
4. Enclosures allow for a simple and automated way to collect extra
content, yes you could run a cron job. But now with a podcatching
client, extra content can be pulled down by less technical people, in my
mind this is a good thing.

As mentioned we (BBC World Service) do use RSS 1.0 to deliver news to
RSS readers, Aggregators and Developers throughout the world.

Post by Bill Kearney
As to processing RDF via XSLT you're in for a whole other set of
adventures...

Can I just ask you to explain this point?

Thanks,

Ian Forrester - Cubicgarden.com



Yahoo! Groups Links

<*> To visit your group on the web, go to:
http://groups.yahoo.com/group/rss-dev/

<*> To unsubscribe from this group, send an email to:
rss-dev-unsubscribe-***@public.gmane.org

<*> Your use of Yahoo! Groups is subject to:
http://docs.yahoo.com/info/terms/

[Bill Kearney]

Post by if
1. Its pretty easy to parse and there are many tools to help

If you're looping through data then it's just data, that it's structured as
items inside an rss element really doesn't make all that much difference.
I'd daresay that it's probably going to be MORE work to cram something into
RSS than to just present it in a relatively direct raw form.

Post by if
2. You can serve 2 or more audiences, so our RSS feed is quite heavy and
contains information which people can read using a RSS reader or parse
and use in there own applications.

Err, so why not provide the RSS audience with it's lighter-weight, properly
formatted data, and leave the interchange stuff separate? If anything,
things like an rdfs:seeAlso URI would be a way to indicate to a "specially
aware" client that additional data is available. That way you're not
pushing out needlessly bulky data to clients that have no use for it.

Post by if
3. Its very common, so many clients now want RSS over anything else (at
least thats true in the news market)

As in "when all you have is a hammer, everything looks like a nail"? Not a
very good excuse, frankly.

Post by if
4. Enclosures allow for a simple and automated way to collect extra
content, yes you could run a cron job. But now with a podcatching
client, extra content can be pulled down by less technical people, in my
mind this is a good thing.

Which makes huge assumptions about how effectively the clients will or won't
handle the content. Not all tools will be as smart about ignoring that
content.

But since you're talking about smarter clients on the retrieval side, ones
that will know how to parse out this 'special data' then why bother jamming
it into an RSS feed anyway?

Post by if

Post by Bill Kearney
As to processing RDF via XSLT you're in for a whole other set of
adventures...

Can I just ask you to explain this point?

RDF in it's full glory can present significant challenges to the limited
ways tools like XSLT are fixated on hierarchical relationships within the
data. STFW for Simon St. Laurent's adventures in handling FoaF RDF as if it
were XML. For a simple document it's probably no big deal but for something
that uses rdf:Description constructs it can be a bit convoluted. Mainly
because XSLT isn't geared to handle RDF's non-hierarchical capabilities.

-Bill Kearney




Yahoo! Groups Links

<*> To visit your group on the web, go to:
http://groups.yahoo.com/group/rss-dev/

<*> To unsubscribe from this group, send an email to:
rss-dev-unsubscribe-***@public.gmane.org

<*> Your use of Yahoo! Groups is subject to:
http://docs.yahoo.com/info/terms/

[if]

So generally you would disagree with RSS as a web API? like how Greg
describes it? [1]

Realistically if you look at the average RSS feed from a blog with full
text content and roughly 10-15 items, you can easily break the 50k mark.
My current RSS 1.0 feed for my blog [2] is 85.07 kB

Clients will be client, is what I tend to say. I've seen clients throw
errors on Dublin core metadata, subsets of escaped HTML in the
description, xml:lang, etc etc. I don't believe this is a good enough
reason to stop adding additional content to a RSS feed. In my
experience, people will change clients if one does not work for them,
specially if you suggest one.

I fear I may be going off topic, but weight and clients should not stand
in the way of a person wanting to syndicate more content than a short
line of descriptive text. Namespaces and RSS extensions are really
useful and give people the chance to build upon well thought out
schema's, instead of reinventing the wheel serving it over REST. Its
reassuring to know that in every service online, there will be a RSS
feed which has a certain base level of content and structure. I believe
this is also why RDF is preferred over plain XML in some circles.

In regards to RDF parsing, I have noticed odd things with RDF parsing
with XSL but nothing major. But the most complex RDF I've had to parse
is FOAF. I'm interested to know more about the problems people face with
RDF parsing in XSL.

Cheers,

Ian Forrester - Cubicgarden.com

[1] - http://www.blogdigger.com/blog/2005/09/07/1126069971000.html
[2] - http://www.cubicgarden.com/blojsom/blog/cubicgarden/?flavor=rdf

Post by Bill Kearney
RDF in it's full glory can present significant challenges to the limited
ways tools like XSLT are fixated on hierarchical relationships within the
data. STFW for Simon St. Laurent's adventures in handling FoaF RDF as if it
were XML. For a simple document it's probably no big deal but for something
that uses rdf:Description constructs it can be a bit convoluted. Mainly
because XSLT isn't geared to handle RDF's non-hierarchical capabilities.
-Bill Kearney
Yahoo! Groups Links

Yahoo! Groups Links

<*> To visit your group on the web, go to:
http://groups.yahoo.com/group/rss-dev/

<*> To unsubscribe from this group, send an email to:
rss-dev-unsubscribe-***@public.gmane.org

<*> Your use of Yahoo! Groups is subject to:
http://docs.yahoo.com/info/terms/

[Bill Kearney]

Post by if
So generally you would disagree with RSS as a web API?

<sigh> The point isn't to argue interpretations. You asked about grafting
non-RSS data into an RSS feed. The suggestion is why bother? If you've got
data that can't be processed by a 'regular' RSS client why bother burdening
it with wasted bandwidth?

Post by if
I fear I may be going off topic, but weight and clients should not stand
in the way of a person wanting to syndicate more content than a short
line of descriptive text.

That's not the initial argument you presented. While, as a premise, there's
validity to the notions of extensibility, there's also the practical
realities of whether the content has the sort of usability one might expect
to be handled in a newsfeed.

Sure, you CAN add anything you like to a feed. But that doesn't mean you
should. As an analogy, you CAN haul 20 tons of dirt with a pickup truck,
but you'd probably be better served using a dump truck. No sense burdening
an ill-equipped vehicle when the task is known to exceed it's abilities.
Likewise, using RSS for hauling extra data smacks of similarly false
economy.

Post by if
instead of reinventing the wheel serving it over REST.

Oh my, ready for that flame war? RSS is perhaps a fine embodiment of REST.

Post by if
In regards to RDF parsing, I have noticed odd things with RDF parsing
with XSL but nothing major. But the most complex RDF I've had to parse
is FOAF.

Precisely. If you talk about wanting to add data to an RSS feed formatted
in RSS-1.0 then you're well on your way toward the complexities of what
RDF/XML serializations require. That is, if you want the data to retain
it's ability to be parsed as graphs.

-Bill Kearney




Yahoo! Groups Links

<*> To visit your group on the web, go to:
http://groups.yahoo.com/group/rss-dev/

<*> To unsubscribe from this group, send an email to:
rss-dev-unsubscribe-***@public.gmane.org

<*> Your use of Yahoo! Groups is subject to:
http://docs.yahoo.com/info/terms/

[g***@public.gmane.org]

On Mon, 31 Oct 2005 06:47:12 -0700, James Holderness

Post by James Holderness
I'd recommend against it. You're assumedly going to have a server at one
end that's not going to be an RSS server, a client at the other end
that's not going to be an RSS client, so why use RSS for the protocol
between the two?

It would be more of a peer-to-peer situation, where each machine would be
both an RSS server and an RSS client, if indeed RSS were to be suitable as
the data transport mechanism.

Post by James Holderness
If you really think the RSS protocol fits your needs perfectly and you're

That would be a point on which I would need some guidance. I need a
transport
protocol, and RSS is that, but it certainly isn't the only one. It just
seemed
more of a match to the content, which would be RDF, but whether or not
that's
important I would leave to the advice of others.

Post by James Holderness
not just stuffing your data into RSS because you don't want to come up
with a format of your own,

I don't want to reinvent anything if I don't have to.

Post by James Holderness
at least give it a different top level element so you don't run the risk
of someone coming across your feed and trying to subscribe to it.

Good point.

Regards,
Brian
--
Using Opera's revolutionary e-mail client: http://www.opera.com/mail/





Yahoo! Groups Links

<*> To visit your group on the web, go to:
http://groups.yahoo.com/group/rss-dev/

<*> To unsubscribe from this group, send an email to:
rss-dev-unsubscribe-***@public.gmane.org

<*> Your use of Yahoo! Groups is subject to:
http://docs.yahoo.com/info/terms/

[Jon Hanna]

Post by bgranam
Greetings.
I've been musing about the possibility of using RSS as a kind of
pseudo Web Service and would like some opinions on the concept.
RSS would be used as a transport layer for arbitrary RDF content
between two or more computers. The RDF content would not be intended
for direct display, but would be parsed out with XSLT and fed into
another program for further processing.

I've never understood the desire to build transport layers on top of
HTTP and call them "Web Services".

TCP/IP is a proven transport layer.

SSL and TLS are proven encrypted transport layers.

HTTP is a proven application layer that works on top of any of those.

Use that application layer to provide a service and you have a web service.

All the rest seems like busywork tbh.



Yahoo! Groups Links

<*> To visit your group on the web, go to:
http://groups.yahoo.com/group/rss-dev/

<*> To unsubscribe from this group, send an email to:
rss-dev-unsubscribe-***@public.gmane.org

<*> Your use of Yahoo! Groups is subject to:
http://docs.yahoo.com/info/terms/

[Bill Kearney]

Post by bgranam
RSS would be used as a transport layer for arbitrary RDF content

RSS isn't a transport layer. It's a data format. One commonly transported
over HTTP. If all you want to do is move data from one machine to another
then using HTTP is perfectly reasonable. Why bother formatting it as RSS if
it's not a newsfeed? There's nothing 'magical' about RSS if it's not for
something like a newsfeed.

-Bill Kearney
Syndic8.com




Yahoo! Groups Links

<*> To visit your group on the web, go to:
http://groups.yahoo.com/group/rss-dev/

<*> To unsubscribe from this group, send an email to:
rss-dev-unsubscribe-***@public.gmane.org

<*> Your use of Yahoo! Groups is subject to:
http://docs.yahoo.com/info/terms/