-
isbnsearch 0.5 released
A distributed search portal of common sources of ISBN numbers, with permanent caching of results. To provide a free software, open-source, interface for ISBN retrieval using HTML, SQL or XML to be independent of any toolkits or software.
Version 0.5 has now been released. v0.5 includes support for the 13digit ISBN format and scripts to recalculate all existing 10 digit ISBN records. isbnsearch scripts will still output the 10 digit form until the 13 digit forms start appearing.
There are also improvements to the matching.php script and the two Perl scripts.
2004-11-02 11:15:38 UTC by codehelpgpg
-
13digit ISBN support in v0.5
With v0.5, isbnsearch has completed the adjustments to the database tables and the scripts to correctly verify the 979 namespace.
Although 13 digit ISBN numbers won't be in use for a while yet, isbnsearch has moved to a 13 digit system by adding the 978 and the appropriate checksum to existing records using v0.5.
All sites should now to upgrade to v0.5. All web pages and scripts will continue to display ISBN values in 10 digit form for some time.
v0.5 includes a single script that recalculates all your existing tables.
2004-11-02 11:11:50 UTC by codehelpgpg
-
isbnsearch prepares for 13 digit ISBN numbers
http://www.collectionscanada.ca/iso/tc46sc9/isbn.htm
The 13digit ISBN is an extension of existing EAN barcodes which isbnsearch already supports. As well as using the existing 978 prefix (plus check digit), the 979 prefix will also be used for ISBN details.
A 13 digit ISBN could therefore start 9-78 or 9-79. All existing ISBN's will remain available as 9-78 and any 10 digit ISBN entered will be converted to 13 digits using 9-78.
To retain the ISBN as a unique reference, newly printed ISBN's will eventually include the 978 or 979 instead of just assuming 978 - in effect, the ISBN is becoming the EAN. isbnsearch will require small adjustments to the database tables and the scripts to correctly verify the 979 namespace and this work is now under way.
Although 13 digit ISBN numbers won't be in use for a while yet, isbnsearch will be moving to a 13 digit system by adding the 978 and the appropriate checksum to existing records using small changes to the database tables.
I anticipate having the basic transfer method in place for the next release of isbnsearch (which will be v0.5). This will enable sites to upgrade the existing tables to 13 digits. This will be a background change *only* - to avoid confusion with the users, all web pages and scripts will continue to display ISBN values in 10 digit form for some time.
Currently, scripts convert any EAN to the 10 digit ISBN. The next release of isbnsearch will reverse this so that all 10 digit ISBN's are converted to 13 digits using 978. v0.5 will include a single script that recalculates all your existing tables.
The new ISBN is a simple mathematical derivation of the old and both are already supported, so each site operator will be able to change over in their own time.
2004-09-08 09:54:50 UTC by codehelpgpg
-
isbnsearch-0.41 released
isbnsearch-0.41
The greatest limitation on any isbnsearch site is the size of the database files. On webhosts, webspace costs money and most hosting accounts offer hundreds of time less storage than a home system. This can translate into
millions of extra ISBN records. To tap into this resource, v0.41 of isbnsearch has been adapted to access systems that are installed behind firewalls and using dynamic or static IP broadband connections. These portals are called 'private' and divided into 'internet-visible' and 'non-visible'.
There are two new files in www/, designed to help an indexserver to communicate with linked portals, other indexservers or private sites. missing.php does not
need configuration.
matching.php is a new script and will be a useful tool in the overall distributed framework for isbnsearch.
Indexservers can now arrange to share data with other indexservers to reduce the number of pending records listed on overloaded indexservers. It should be called
from your local indexserver and can query all pending records of the chosen remote indexserver, 1,000 titles at a time. Any records on your site that match pending records on the remote site will be written to a temporary XML file,
temp/missing.xml. matching.php then offers a WWW link to the chosen indexserver to allow collection of the XML file. You will see the results of the import and be provided with a suitable link back to your own indexserver with the next 1,000 records. It is possible to automate this process using GNU/Linux utilities like wget - please report all experiences to the mailing list.
The chosen indexserver MUST be configured to accept data from your site. Operators must request this service from the maintainer of the chosen indexserver using the mailing list by giving the full URL of their site. The indexserver maintainer can then verify the availability of your site - if this fails or is unacceptably slow, the indexserver maintainer is free to decline the request - and add appropriate details to the configuration of private.php on the remote indexserver. Without this permission, you can run the match against the pending titles, but the remote indexserver will not import the data - the temp/missing.xml file will simply be overwritten.
An 'internet visible' private portal has a static IP and a valid DNS name, the indexserver must be able to resolve the name and retrieve a file from the private portal using HTTP on port 80. It is NOT necessary for HTTP on the
private portal to provide any service to the general internet. Any private portal that cannot provide a local file over the internet in this way (dynamic IP, firewalls etc.) is deemed 'non-visible' for the purposes of this section.
2004-07-12 10:24:45 UTC by codehelpgpg
-
Development continues
After a break to deal with other projects, I've been spending time again with the isbnsearch code. The next release is set to include more code for private sites to participate. These sites offer very large storage capacity and could be a useful resource for heavily loaded isbnsearch indexservers. The main code has had a few tweaks and current CVS HEAD will be folded into the next release, 0.41, in the next few weeks.
2004-07-03 23:26:14 UTC by codehelpgpg
