waldo at vqronline
Apr 17, 2008, 12:11 PM
Post #8 of 8
(377 views)
Permalink
|
On Apr 17, 2008, at 2:38 PM, David E. Wheeler wrote:
> On Apr 17, 2008, at 11:30, Waldo Jaquith wrote:
>> There are a few reasons. The first is that it wouldn't make sense
>> within the scope of our directory structure. We keep all of the
>> content for all issues in /articles/year/season/slug/. It's simply
>> illogical to present it as /articles/category/year/season/slug/ or
>> as /category/year/season/slug/. (It implies a division of our
>> content, site-wide, by the genre of the work. That's not a division
>> that we make or care to make.)
>
> In that case, I personally wouldn't use /articles. I'd just put
> everything in the root category, so you got /year/season/slug. /
> articles/ is just redundant.
The only reason I'm using /articles/ now is that our existing CMS
requires *something* that stands between the articles hierarchy and
the root. Because, yeah, /articles/ makes me cringe, too. :) That may
be the one change that I make during this transition.
>> The third is that the URLs would become entirely too long. And the
>> fourth is that the additional information, while more descriptive,
>> wouldn't add any substantially useful information for our readers.
>
> It's not for your readers, it's for the search engine (although I
> personally get a lot out of descriptive URLs, but I'm a geek).
What I should have said that it wouldn't add any substantially useful
information *that would be worth the longer URLs*. (Thus conflating
points 3 and 4...but it's so *nice* to have a whopping four points to
make. :)
I have to wonder how useful that genre data is to search engines.
Surely they don't treat nonfiction and fiction separately. And I don't
know that folks searching for "fiction," in general, are ever going to
find our publication appearing prominently, no matter how awesome our
Google-fu is. :) This is the kind of thing that Dublin Core is really
meant for...though, thanks to spammers, site-defined metadata is
basically useless. OTOH, providing more metadata to search engines (or
Greasemonkey, or whatever) is *always* good. Hmm.
>> Simon and Chris, thank you very much for your suggestions. In
>> particular, Chris, it hadn't crossed my mind that keywords could be
>> useful for this purpose, but it makes perfect sense. I'm not sure
>> which of the two approaches that I'll take, but in importing all of
>> my existing content, I believe I'll allow for each possibility.
>
> I think that key words are better for this, unless you're also using
> them for something else (like keywords for your documents), in which
> case I'd go with the secondary category solution (which I first did
> for Macworld back in the day).
That is a bit of a conundrum for me. Keywords are an awfully useful
thing, for using internally, and I'd hate to give that up. OTOH, it
seems like a fantastically useful thing, to just generate categories
from keywords. Bricolage, like Perl, seems to have a real strength in
providing multiple methods of accomplishing the same thing, so I'll
figure something out. :)
Best,
Waldo
---
Virginia Quarterly Review
One West Range, Box 400223
University of Virginia
Charlottesville, VA 22904-4223
|
