[Up: Weblogs FAQ] [Robot Wisdom Weblog]
Some sites give you long, messy URLs that are hard to send via email, impossible to share over the phone or on a post-it, and ugly to post visibly in a weblog.
Often this is because they're generated by an HTML FORM, which uses this format:
http://www.site.com/directories/stub? a=1& b=true& c=%22a+string+of+characters%22
Usually you can trim away most of that and rely on defaults.
Sometimes (eg Amazon) the site is using the URL to create a cookie or 'shopping cart' that lets them track where you came from and which of their pages you're looking at.
Deja.com
Deja.com gives you URLs like:
http://x24.deja.com/viewthread.xp? AN=583874412& search=thread& svcclass=dncurrent& ST=PS& CONTEXT=951202142.2136801298& HIT_CONTEXT=951202142.2136801298& HIT_NUM=2& recnum=%3c87srra$qki$1@bob.news.rcn.net%3e%231/1& group=alt.cars& frpage=getdoc.xp& back=clarinet
Happily, you can trim this to:
http://x24.deja.com/viewthread.xp? AN=583874412
That one's linking to a thread, here's a single message:
http://x24.deja.com/getdoc.xp?AN=583874412.1&CONTEXT=951202142.2136801298&hitnum=2
becomes
http://www.deja.com/=dnc/getdoc.xp?AN=583874412
(The "=dnc/" segment gives the readable 'Deja News Classic" format. The ".1" at the end of the Article Number (AN) tells Deja to break it into short segments and show the first, which is usually very annoying. Changing x24.deja.com to www.deja.com is probably good practice.)
Amazon
With Amazon you can ruthlessly strip everything back to the first long number. So:
http://www.amazon.com/exec/obidos/ASIN/0439139597/ref=gw_m_r_bo_1/102-6315101-2367217
becomes:
http://www.amazon.com/exec/obidos/ASIN/0439139597/
And you should always make sure someone worthier than Jeff Bezos gets the referral fee, by pasting the relevant ID at the end:
http://www.amazon.com/exec/obidos/ASIN/0439139597/progressiverevieA/
NY Times
There's a very simple trick that saves people having to register to read articles on the NY Times site. Instead of
http://www.nytimes.com/books/00/05/28/home/contents.html
just substitute 'partners' for 'www':
http://partners.nytimes.com/books/00/05/28/home/contents.html
Excite maps
Excite's map-service gives fast, clear maps with a very high degree of customzability. Here's a Joyce-related example, that shows Dieppe (France), Paris, and Newhaven (England). Newhaven is too small to show up at this scale, but Excite allows you to specify it as a label if you know what you're doing.
The critical fields are:
http://maps.excite.com/zoom/? lat=499333& lon=10833& level=4& orig_lat=507833& orig_lon=333& label=Newhaven
(If you don't need the label you can drop the last three.)
Start from their search form.
If the first result is good enough, you can just use
http://maps.excite.com/address_results/?c=newhaven&cn=england
Irish Times archive
When you search some archives like the Irish Times, they give you a result that highlights your search term in red (or bold). This is done via the URL:
http://scripts.ireland.com/search/highlight.plx?TextRes=livia&Path=/newspaper/ireland/1999/0728/courts7.htm
And in this case can be trimmed back to the elegant:
http://www.ireland.com/newspaper/ireland/1999/0728/courts7.htm
The same principles can be used to copy and customize searchforms for your own pages. This Google form:
looks like this if you 'View Source':
<form action="http://www.google.com/search" method=get name=f> <input type=hidden name=num value=20> <input type=text name=q size=40 value=""> <input type=submit value="Search"></FORM>
You can simplify your version by pre-setting your own defaults with "type=hidden" inputs.
Taking this a step further you can make a search-bookmarklet like Google that squeezes it all down to a single link (requires Javascript).
[Up: weblogs FAQ] [Site map] [Robot Wisdom home page] (Feedback)