Roy Schestowitz <newsgroups@xxxxxxxxxxxxxxx> wrote:
> __/ [John Bokma] on Wednesday 07 December 2005 06:00 \__
>
>> Perl script that automatically creates a feed:
>> http://johnbokma.com/perl/rss-web-feed-builder.html
>>
>> feedback (you can post on my site) and back links are welcome (as
>> always).
>
> Okay, I was seriously about to set this up. I wanted a
> recently-added static pages feed (I notice that all of yours a
> static).
>
> A random few thoughts:
>
> -With large sites, the scripts hogs resources for a very long time,
> which becomes problematic on a shared server. 'nice'ing it might help.
> I suppose one could use this on small sites with fewer worries.
> Advanced stats pack- ages are just about as resource-greedy.
I will look into that.
> -Although I can install the XML-RSS Perl module, I cannot install it
> on my host's machines. I can use phpshell to get shell access, but no
> root priv- ileges.
Ah, but you can install the module locally. I will add a how to.
> -The script is a neat one to put on the Web server and define it to be
> run nightly using the crontab.
Yup, that was one of the ideas, either that, or run it on the local
machine. My site is created on my local machine, and only new pages are
uploaded, which might include the index.rss.
> Again, there is the issue of
> scalability. cPanel allows cron jobs to be defined without the need
> for workarounds such as shell/SSH/telnet 'deprivation', which is
> often set by default.
>
> -The script seems very customisable and simple enough to tailor to
> one's needs.
Yup. I am open for suggestions :-D.
> -if the script cannot be automated to run independently, it's
> almost pointless.
Not at all, see above. If one has a local copy of the website, it can be
run locally. However, this means that one shouldn't have some program
that re-generates all files on the website no matter if they have new
contents or not.
> I noticed that you periodically modify your XML
> files, which makes update non-real-time.
I run a mkfeed.pl script manually. I make my site as follows:
I have a bunch of XML files, and a perl script that parses those files,
and generates HTML files from it (not 1:1, one XML file can generate
several HTML files). The generation of HTML is done in memory. For each
HTML page, the in memory version is compared with a version of a
previous run, if there is a change, the in-memory version is written to
disk. So when the program finishes, it gives a nice overview of which
pages are new.
A second script compares each html file in the web directory with a copy
of the actual website. If there is a change, the new file is uploaded,
and added to the local copy :-D. So only files that are new/changed are
uploaded.
The mkfeed script just parses a text file, and creates a feed based on
the info in it. What I do is I add an URL to the end of the file for
each new item. The script extracts the info similar to the feed-builder
script, and creates the feed :-)
> All in all, the script would suit many people, but it does not
> accommodate my circumstances.
And the second one? (i.e. you copy URLs to a txt file)
> I guess I could run the Perl
> from here ( http://baine.smb.man.ac.uk:8001 ), but putting my
> feeds off-site would be similar to 'offshoring' feeds to FeedBurner.
> it also devours bandwidth.
The main reason is XML::RSS or the resource usage when the HTML files
are scanned?
> Roy
>
> PS - I like scripts that are practical such as ones that serve as SEO
> tools and automated site 'housekeeping'. Keep em' coming!
Thanks, suggestions are very welcome for new tools :-D
--
John Perl SEO tools: http://johnbokma.com/perl/
or have them custom made
Experienced (web) developer: http://castleamber.com/
|
|