isbnsearch-devel - Home

Purpose: To develop and maintain isbnsearch.

• Full documentation for isbnsearch is available, including:
  • Install notes,
  • upgrade instructions for users of previous releases
  • and details of how the isbnsearch project is structured.
• For support or queries, please join the isbnsearch-devel mailing list.
• For general information on the project, please see the main SourceForge pages.
• For a look at how the project website operates, see the active site.
• For information on why certain ISBN records might not be found, please see the FAQ.

On this site

RSS Feed: Project news releases (including full text of news items)
Other feeds are available from the SourceForge site.

• This is the home of the development section of the isbnsearch project. Development files are updated with each isbnsearch package release. All development files are available for download from this site and from the main isbnsearch devel SourceForge site. All devel files are also available in the isbnsearch/devel/ folder of the main isbnsearch CVS tree held at SourceForge. Development files are no longer included in the main isbnsearch release, only in the development release.
• Verification notes and scripts from isbnsearch.
  • Readme.
  • Preparation.
  • Verification.
  • Verify.pl.
  • verified Z39.50 servers report.
  • Z39.50 servers that fail verification.
• All servers are re-tested after major code revisions and before each package release. Changes will be updated on these pages. Last update: v0.5 release. Wed Nov 3 11:41:37 GMT 2004
• Servers listed on the LoC gateway but that failed to initialise the Z39.50 protocol during tests using the Library of Congress WWW/Z39.50 Gateway interface are ignored. Other servers listed as FAIL may have accepted a normal Perl socket request on the Z39.50 port but failed to complete a Z39.50 connection on the same port. This may be due to configuration faults on the server or firewalls. isbnsearch tries to use canonical names in preference to IP addresses for two reasons:
  1. IP addresses generally have a shorter life than the canonical name
  2. A DNS failure on a canonical name is easier to detect than a failed IP address that may or may not have a reverse-lookup name and will usually not respond to ping requests but will often allow a Perl socket request to complete, even if the service on that port is down or disabled.
• Canonical names that fail to resolve via DNS are ignored and are no longer even listed in the server reports.
• The servers listed on this site are a combination of LoC Gateway servers and Net::Z3950::AsyncZ servers with any canonical names tested in DNS. Every server in the LoC Gateway page or the AsyncZ DB.txt list on Tue Nov 2 2004 has been tested.

---