Do you have a question? Post it now! No Registration Necessary. Now with pictures!
- Posted on
- RSS: Assorted Ignorant Newbie Questions
October 16, 2006, 2:37 pm
rate this thread
I've read an introductory article on wikipedia about RSS. I still
have a collection of ignorant newbie questions. I don't mind being
redirected to a *good* source to read the answers myself. Here goes:
1. To put another sites RSS feed on my own web site, do I need my ISP
to do anything for me or do I just put all of the appropriate tags in
on my site?
2. To set up RSS for others to use, do I need anything from my ISP?
3. The wikipedia article made it sound as if Atom feeds were the
better deal. It Atom the most popular. If I made feeds available
only in Atom would I be inconveniencing anyone?
4. Is RSS a pull or a push technology?
5. Can various RSS feeds be combined into set of output on a page?
Would that be done by putting it all into a database and then recalling
Thanks in advance for any info
Re: RSS: Assorted Ignorant Newbie Questions
There are lots of ways to do this. In general you need some support for
scripting on your site, although this is unlikely to require your ISP
to do anything for you (and it's even less likely they'll agree to).
The amount of scripting depends on the sophistication you aim at,
modulo what's available on your ISP. There are also ways to do it
without any server-side scripting at all.
A typical "good practice" approach is for your server to have a list of
"inbound RSS feeds" and for it to go and read them at appropriate
intervals. Their content is the stored in a database. This content is
then served up as either "outbound" RSS feeds, or as HTML.
Transformation to HTML is probably done via XSLT. The results of the
transform are cached, so that it can be served quickly on-demand,
rather than waiting while it's re-generated.
You can write all this, or download other peoples' versions. In proper
languages, or even in PHP.
A simpler version of this (and a thoroughly bad idea) is that when a
user of your site wants to view the RSS feed as HTML, your server
downloads the RSS, transforms it with XSLT, then serves the results
back to the user. This has the following problems:
* It's slow. Lots of work is done at the time-critical part. Your user
has to wait.
* It's unreliable. What if the RSS feed can't be retrieved at that
time? You can't serve the results to your user.
* It's inefficient. Every time you have a user, you hammer the upstream
server. If you're busy, this sort of behaviour is likely to get your IP
blocked for greediness.
Doing it without server-side scripting can be done with <iframe>s, with
Many RSS feeds include a client-side XSLT stylesheet (the BBC news site
is a good example). If you load this from a plain and simple web
browser that groks XSLT (many modern ones), then you get a client-side
transform of it into usable HTML. Some RSS servers will even do this
for you, if you approach them with a HTTP Accept header that looks more
like a bare browser than an RSS-collector. If you take such a feed an
embed it into an <iframe> (i.e. the src attribute points straight at
the external site) then it will "work", for an <iframe>'d page that
loads a 3rd party's server and has 2/3rd of the disadvantages listed
above. Don't do this unless you have a pre-arranged agreement with the
cross-server limits in techniques like SSI. You can't use SSI between
document happens to be a program that writes HTML content into your
document that happens to be the results of an RSS feed transformed to
HTML, then you have embedded an RSS feed onto your page. There are
services out there that will do this for you, for RSS feeds you point
them at. I don't like this, but grudgingly admit that it works - for
sites that can't do any better. You're beholden to the RSS -> JS
converter site and it also depends on your users having JS enabled. It
may also give dire warnings (especially under IE) about 3rd part JS
attacks and the fearful and na=EFve might think that this is a phishing
It's also necessary to ask why it's useful to put an external RSS feed
onto your site, and why it's useful to do this for a non-aggregated
feed. Link to external feeds by all means, but ask yourself why I (as a
website user of your site) should choose to read an RSS feed via your
site, rather than going directly to the feed and doing it there? What
are you _adding_ ? What extra value are you giving me, the website
For some sites, external news headlines are a good idea. Not many
"hobbyist" sites though.
Some sites are themselves RSS aggregators. I read most of my RSS these
days through LiveJournal, rather than through my RSS aggregator.
No, this is really easy. There are a few ways, broadly "static" and
For "static", then write some RSS and upload it as a plain old static
file. You might need the web server to be configured with an
appropriate HTTP content-type, or a clueful admin will have already
done this. You can fix it yourself by editing .htaccess, or by asking
the admins nicely. As it's a good idea and generally useful to
everyone, then an admin ought to be well inclined to do this (if they
won't, then leave that ISP).
How you generate it is up to you - but it should be up-to-date,
accurately encoded, and accurately reflecting your HTML web content.
I have a number of small catalogue sites up and running where I
literally author the site content as an RSS document (by hand) then use
XSLT to generate the site HTML from it (i.e. I make the site from the
RSS, not the usual other way round). I can make a 50 item catalogue
this way, just by editing a single terse RSS document. Geeky, but it
I also have a script that scrapes some of my HTML sites, reads metadata
from the pages embedded in <meta> tags (even if Google doesn't trust
them, I still use them and find them valuable just for this reason).
A "dynamic" RSS feed is based on some content-management system,
probably based on a database. As well as the query that makes the HTML
pages, another query (much simpler) makes an RSS version.
There are also the "offline batch" techniques for building sites. An
offline dynamic CMS on your desktop is used to build (i.e. export) a
new site from time to time, then it ftps it onto the web server.
RSS 2.0 has biggest market share.
RSS 2.0 is pretty much essential if you're targetting the podcast
RSS 1.0 is technically the best for some complex needs, but if you're
not going to use them, then it's really not the format to choose.
Atom is the best all-around. RSS aggregators will accept it just as
well as any other version. Some crude RSS tools won't, but then they're
not a significant market.
Many RSS server scripts support multiple formats automatically,
depending on what the user wants.
I'd recommend Atom
Both. Like a lot of web technologies it's fundamentally pull, with
tweaks to make it act like the push that we'd actually prefer, but
can't practically implement.
In addition RSS does also have some pure push features added to it
(look at "clouds" in RSS 2.0)
This is called "aggregation".
There are three sorts of aggregator: simple, de-duplicating, and
Simple aggregators take a number of RSS feeds and combine their items
into a single result. This result might be in a different RSS version,
or might wrap them up with advertising / mangle them in some other way.
Simple aggregators have a problem in that many users will be
interested in a number of closely related RSS feeds. If one feed
syndicates a related article to the others, then the aggregator will
return duplicates of it.
A de-duplicating aggregator fixes this. It can recognise items by GUIDs
or permalink values, then strip obvious duplicates. It might even do
text comparisons and filter likely duplicates.
Both of these return a list of items that's basically the union of all
the items in all the fields. That's a lot of content....
A filtering aggregator is more interesting. It "knows" what you want,
and reduces the feeds to just those items. You can point one at a _lot_
of busy feeds and just read the highlights. In practice it's very hard
to make them work - they rely on metadata about the item content so
as to judge if it's relevant. This can either be attached to the <item>
by its author (best, but rare) or it can be inferred by text analysis
of it (as the text is short, this is difficult to do well).
Re: RSS: Assorted Ignorant Newbie Questions
Amen. I've seen a lot of pointless along these lines myself. In my
area there are people who want to get event news from about 4 - 5 orgs,
each with their own web stie for good reasons, and about 1/2 dozen
individuals. Right now people get news of these events with a half
dozen or so email lists. It is pain for everyone involved. I would
like to put up my own site that would collect all of this information
in one place and sort it. I was thinking RSS would be ideal. Nobody
would have to subscribe to a list and the event organizers for these
orgs wouldn't have to do anything. Once their announcements go up on
their site, it is on my site. I have good enough relations with these
orgs that I can get cooperation.
I've been experiementing with RSS feeds using Sage for Firefox and
getting stuff on MyYahoo and Google pages. I notice that the feeds
seem to disappear for a while and then come back. What is that about?
Some of the feeds are from sites that require authentication. Is
authentication an issue with RSS? Are there better places to get a
customized RSS feed page ( & why)?
What would be a good RSS referrence for me to start with given my
My ISP has fantastico and Java. Given that I want to aggregate several
feeds, sort them, and have some system for getting input from
individuals what would be the best route? Installing a content
management system and then customizing it?
Thanks in advance for any pointers
- » shock & awe: rhode island school of design web site
- — Previous thread in » HTML Authoring Forum