Yeah, I over-simplifed and mis-stated.  My current code, against a weird 
IA admin form that is scrapable and useable as an API even though not 
intended as such -- does use author/title text, in fact, not just 
ISBNs.  In fact, it can't use ISBNs even when I'd like to.

The current code has been in production for a few years, and works. But 
I'd kind of like to switch it over to the actual supported OL API.

But I kind of want to know how many full text books I might be losing 
access to if I make this switch, because they are in IA but not OL.

This is reasonable, yes?

[My code works from a 'known item', and tries to find matches in IA/OL.  
When I have an ISBN, I'd like to use it, to avoid false positives. Just 
author/title gets you lots of false positives. But in fact my current 
code doesn't do this, it always uses author/title. And an OL API 
implementation could do that too if desired. So this is really an 
orthogonal question. ]

On 2/2/2011 12:17 PM, Lee Passey wrote:
> On Wed, February 2, 2011 8:29 am, Jonathan Rochkind wrote:
>
>> Can you give us a broad overview of how many IA books with full text
>> have marc records, and how many don't, and what sorts of sources don't
>> have marc records?
> This would be interesting data.
>
>> I have some old (pre-OL) code that searches for IA full text via ISBN
>> through a kind of crufty means. I am considering switching it over to
>> the OL api, but want to know how many hits I might be missing if I do that.
> While I cannot speak for the OL catalog team, I would guess that searching for
> IA text using the OL api would be much more effective than trying to use ISBN.
> Why? ISBNs didn't come into existence until 1966, and the vast majority of
> texts at IA (at least those accessible to the general public) will be those in
> the public domain, i.e. those published before 1923. If you try to search via
> ISBN you will get, at best, modern reprints of the classics.
>
> ISBNs might be a good way to find /catalog data/ about /modern/ books, but
> they are unlikely to lead you to full text (actually, the full text at IA is
> virtually unusable, but at least it would get you to the page scans).
> _______________________________________________
> Ol-tech mailing list
> [email protected]
> http://mail.archive.org/cgi-bin/mailman/listinfo/ol-tech
> To unsubscribe from this mailing list, send email to 
> [email protected]
_______________________________________________
Ol-tech mailing list
[email protected]
http://mail.archive.org/cgi-bin/mailman/listinfo/ol-tech
To unsubscribe from this mailing list, send email to 
[email protected]

Reply via email to