Feb 042011
 Posted by on UTC 2011.02.04Fri at 15:35 1.JotHere.com admin, web link admin LG49K3  Add comments

LG464P: :  (~50% complete)

  1. LG46H3: Web URLs being broken-else-repairable is a deep subtle problem that often is overlooked when one is first putting content up but can prevent that content from scaling as any change breaks too many links.
    1. LG46LZ: Many of my past websites grew to where one couldn’t move anything without all the existing links being broken, so even when it needed to be moved it wasn’t, else the maintenance of links became an overwhelming task, killing the scalability.
  2. LG48EX: Comparison of possible solutions :

    1. LG48HD: Comparison table: (to be constructed)
  3. LG46OW: Here are possible solutions, non-exclusive (one use any number or all of these):

    1. LG43RK: HTTP redirection: have the old location redirect to the new location

      1. LG47RV: The obvious solution.
      2. LG471P: It is built-into the HTTP protocol and some & increasing websites support it.
      3. G441L: This method has significant advantages in that:
        1. LG47MC: It is easy to automate in forwarding & virtually instantaneous
        2. LG47MT: It has methods to direct that links be updated though my guess is these updates aren’t often done.
        3. LG47OZ: It assures the accessor is getting the right redirection location (not tricked by phishing)
      4. LG4718: Requires that prior location have all the following:
        1. LG4754: Be available including:
          1. LG478P: still hosted.
            1. LG47A3: This is usually not a problem today as hosting can be had for fairly cheap, and all we need is redirection system which (for small loads) can be had fairly cheap.
          2. LG47EV: Not used for some other purpose
            1. LG47IG: Not always easy:
              1. LG47J2: The domain could be sold to someone else who has some other use in mind
              2. LG47SS: One often wants to move one’s website to a different back-end (from Drupal to WordPress) but keep the same domain name, however the new back-end may lack the means of properly processing the old back-end’s paths.
                1. LGWRPL: See site implementation/version ID for a solution.
        2. LG4762: Be cooperative
          1. LG47FG: Not always easy:
            1. LG47FX: The prior platform may be expensive and NOT doing general redirection, but the host/domain owner may not want to replace/add-to it with a  redirection system and just charge for that.
        3. LG476C: Have the ability to do this, including the ability to redirect hundreds or thousands of pages in bulk.
          1. LG4775: WordPress can do redirection with PrettyLinks and a few other plugins, but I’m not sure if much support exists yet for specifying the redirection in bulk.
    2. LG43UJ: Store the OLD URL(s) in the content at its new location (as “former locations” — possibly meta data) and use search engines to find the new location

      1. LG48J6: This is my invention though others may have thought of it, too.
      2. to be continued
    3. LG48L7: Put anchor(s) in the content which are globally unique IDs and, when a link to the content includes one of these anchors, search for that anchor via search engine to find the content (or at least the part with that anchor).

      1. LG48U1: This is my invention since ~2008 or earlier.
    4. LG46Q3: Have the path to the page (else final page name/ID) be a somewhat globally unique ID which stays the same even if the page is moved to a different site.

      1. LG47QZ: This is my invention.
      2. LG46VJ: Usages in practice, all using a universal ID:
        1. LG46VV: http://2.loverules.info/lfnj0t#LFNJ5D
        2. LG46WJ: as much as was possible on http://blogger.LoveRules.Info , for instance http://blogger.loverules.info/2010/05/l2pfgy.html the file name was set to use the GID, the other part of the URL were required by Blogger.
      1. LG42PL: Why do this?
        1. LG4A05: If the post is moved to another site, it has the same path
          1. LG4A12: one could do this if
            1. LG4A0A: (likely) that path wasn’t already in use and
            2. LG4A0M: (semi-likely) the new host site allowed this (to pick a particular path) (WordPress generally does, though other sites often don’t)
          2. LG4A1B: one would do this as:
            1. LG431Q: significant-pro: If the content has been hosted somewhere else over which we have  search engine (if it’s public, then say Google Search) which can search for paths, then any broken link could be automatically or semi-automatically repaired via searching for pages with that path, even if the former site no longer was hosted.
              1. LG43PY: Note there are other possible solutions
            2. LG42Z9: small pro: In content moved from the old site to the new site, if it’s links used a relative reference (as they may typically) then they would not be broken as that reference would be the same.
              1. LG43K7: This is only small pro because
                1. LG43KB: this doesn’t at all solve the tough issue: fixing the external links from breaking
                2. LG43KM: where this would save links from being broken, we could probably readily fix them anyway as if we are moving content we can probably edit it