>>>>> "IT" == Iain Truskett <perl@dellah.anu.edu.au> writes:

  >> so why not have 3 links? same thing as one but you store 3 url's in
  >> the DB (if the book has all 3 sizes).

  IT> Because the site may only have one version? Amazon usually has three
  IT> different versions (it's where all but about two of the covers on the
  IT> current site come from [those two being 'Practical mod_perl', and
  IT> 'Embedding Perl in HTML with Mason']). Or, the site may have one version
  IT> and it's shite. It's usually the author's sites that have the best
  IT> versions (and, yes, we could link there), but ideally all our iamges are
  IT> around about the same size (aspect ratio notwithstanding).

  IT> *Ideally* we have a nice big hires copy that we can scale however we
  IT> like. We, obviously, can't just reference one image and use <img
  IT> width/height> unless we also store the aspect ratio.

  >> > We could always cache and Magick them, but then we'd be storing =)

  >> my whole point. why store when we don't have to.

  IT> Flexibility.

i wasn't thinking about image sizing on the fly. i assumed that if there
was only 1 image, then all 3 urls would use it. much simpler than
worrying about what size it was and making 3 properly sized images from
it. but then i don't do much image hacking. this seems more a nice touch
than a needed feature. i would rather you spend your time on getting the
other links (home page, contents, etc.) going and then we can brainstorm

  >> > There's another kettle of fish. Some places don't even like links to
  >> > images from pages that aren't on their servers.

  >> well, we can deal with that later. this is a non-profit info site. we
  >> can ask the major publishers for permission (o'reilly and manning will
  >> surely let us. i bet most of the others will too).

  IT> That's what I suspect. I suspect they would even be friendly to us
  IT> having copies of the covers ourselves.

yep.

  >> i can do it for those two as i know people there. i don't know the
  >> other publishers. i think we could just do the linking and let them
  >> complain first. then we get permission or drop them.

  IT> Which sounds a good step. We'd have two for precedent =)

ok, i will do that soon. i expect no issues from those two.

  >> it is free publicity (unless the book sux) so they probably won't
  >> mind. out site should help sell books which is what they want.

  IT> Even bad books get good reviews. See Amazon =) The trick with this site
  IT> is to get the Perl community rating and reviewing, because they
  IT> theoretically know better than Joe Dredd who just picked up his Perl/CGI
  IT> book and is looking for something more. And good books get bad reviews,
  IT> etc.

oh, i agree. but we want 2 tiers of reviews, the unwashed masses (anyone
can sign on and add ratings/reviews) and the cabal (those who supposedly
know perl).

  IT> O'Reilly and Manning, due to having ebook editions, would probably love
  IT> the site =) After all, who visits Amazon, looking for a book and then
  IT> finds Safari?

true. and we should have a link to those ebook pages for each book as
well.

my thought is to have some sort of table called book-links with the book
id, the link type (home page, contents, review, etc.) name (reviewer?)
and the url. that way we can easily add more link types and such without
adding fields. ebooks being a new link type and external reviews being
another.

all the book links will be extracted and binned according to type in the
mason code. so we can have the home page, contents, etc in one section,
and the reviews in another. thoughts?

uri

-- 
Uri Guttman  ------  uri@stemsystems.com  -------- http://www.stemsystems.com
----- Stem and Perl Development, Systems Architecture, Design and Coding ----
Search or Offer Perl Jobs  ----------------------------  http://jobs.perl.org