System update

Published Monday, 30 June 2003 12:35AM CST by in Announcements

3

I’ve updated the system to pMachine 2.3 which is a much less conservative update than the version number would indicate. The new features that have already been implemented here include trackback, pingback, multiple RSS feeds, and mailing list improvements.

My RSS feeds (yes plural) are funky because I insist on using typographically correct punctuation elements. Since I do most of my writing these days in Microsoft Word and export to text, the typographic elements come with and I’m too lazy to convert them to their proper named- or numeric-entities.

One of the niftiest features of pMachine is auto-XHTML which, if I dink around with Word’s encoding—or find a text editor for Windows as good as BBEdit for the Mac—will allow me to get standards-compliant XHTML into the pMachine entries. But I’m lazy and it’s an extra step and did I mention I’m lazy? Oh, what the hell, I’ve got a few minutes so I’ll test it with this post. If all goes well—and it should—my typography fetish will be addressed and I’ll have compliant code.

So, here are my funky RSS feeds (any XML parser worth beans will bitch and moan about invalid characters, but every aggregator I’ve tried takes mercy on me and seems to work fine).

I’ll put up an icon and link for the RSS 1.0 (RDF 1.0) feed as soon as I can re-locate that spiffy CSS button generator gizmo (if you know what I’m talking about and know where it is, kindly drop me an email).

Next up will be the new caching mechanism. As soon as I can figure it out I’ll implement it and performance here should be noticeably improved.

Update: It appears that an em-dash must have a single space on each side of it for the two dashes to work. That sucks but is probably fixable. And I had the permissions set wrong on the RSS feed XML files (fixed and should be working now).

3 responses. Comments closed for this article.

  1. Lou Quillio says:
    05 July 2003 at 3:48 pm

    It appears that an em-dash must have a single space on each side of it for the two dashes to work.

    Check the script file pmcode.fns.php, line 380, the function named function xhtml_typography().

    Within it, starting at line 410, a translation-table array is declared.  Check line 425:

    [b]" -- "    => "—",[/b] 

    Change it to:

    [b]"--"    => "—",[/b] 

    Probably want to uncomment line 424 as well, which will transform “...” (three periods) to “…” (Unicode ellipsis, …).

    There’s more hacking to be done to this function, depending on how you like to work.  Though it pertains to pMachine 2.2.1, a longer discussion is here.

    Both of your RSS feeds are broken in Amphetadesk 0.93 and Wildgrape NewsDesk 1.1.8.0, for the reason you cite.  You’re not alone; it’s one of the reasons I haven’t upgraded yet.  Another is that my pMachine installation is hacked within an inch of its life.

    Nonetheless, it seems that pM 2.3’s RSS Factory is not ready for prime time, yet digging into it when the syndication world is in such flux seems like a waste.

    As a temporary fix for folks who compose in MS Word before pasting into pM, I just now threw together a quick VB macro that transforms curly quotes (single and double), em dashes and ellipses to their Unicode representations.  Somehow I’d never fooled with macros for MS Office, but it’s pretty simple.  This one can be easily be altered or extended, and it’d probably be useful to have a macro that “undoes” these transformations.

    Run the macro just before copying your finished post from MS Word.  Get it here:

    unicode1.bas.txt

    Be sure to remove the “.txt” extension.

  2. Lou Quillio says:
    12 July 2003 at 5:52 pm

    Found the characters that are breaking your feed.  In the post entitled “The MoveOn primary” there are three occurrences of ASCII 146, right single quote, which evidently are copy-paste artifacts.  When I changed them to &rsquo; the feed was fixed.  (Actually there may be more in later <item>s in your current feed; I just tested items individually until I found the first offender.)

    Yours isn’t the only feed that suffers this problem from time to time.  Several that I subscribe to go dark periodically, then come back to life when the recalcitrant <item> falls off the feed.  Evidently these extended ASCII characters are the typical reason, and copy-pasting the typical cause.

    One more thing:  &rsquo; (for example) will preserve the validity of the feed but kills well-formedness.  According to the XML spec, numeric entities are required for all but five extended characters.

    Tipped-off pMachine about this.  Either there’ll be a fix issued or I’ll write one.  It’s not that tough to filter this stuff out, at least for the RSS.  In the meantime, the macro above will catch the commonest character errors.  Watch out for fractions, trademark and copyright symbols, exponents, etc.

  3. Michael Fraase says:
    18 July 2003 at 1:38 pm

    Thanks for the heads-up, Lou. This is all so much more difficult than it needs to be/should be for non-technical people. I’m fairly technical, although not a programmer, and it makes me crazy. Non-technical folks—the folks we need to really make nanopublishing zoom—simply won’t play in this kind of environment.

    I knew my RSS/RDF feeds were hokey, but I didn’t know they were dying in some of the aggregators. Maybe the aggregators should be a bit more tolerant? I know my own feeds—and those of other “problemmatic” sites (like all of the Corante feeds)—come through in the aggregator I use. They don’t come through as fully valid nor well-formed, but they come through.

    I think I’ve got a solid workflow in place now: write/edit in MS Word and then save as text with the “allow character substitution” option turned on. The resulting text is then valid and well-formed and pMachine subsequently does an admirable job of re-rendering the copy with typographically-correct elements.

    I’ve only been asking for this since 1993.