CERN documents on the Web
 
EUROPEAN ORGANIZATION FOR NUCLEAR RESEARCH
 
CERN-IT-US-2000-015
January 15th, 2000
Last rev: March 13, 2000

 

CERN document names on the Web

Maria Dimou , Miguel Marquina, Andreas Wagner - CERN/ User Support

This is a proposal for some standarisation in the naming of URLs for CERN internal documents. The aim is to:

  • help authors to publish reports, minutes of meetings, group documents etc
  • help users to find them.

The scope:

The documents produced in IT are the first customers but the naming and registration scheme covers CERN-wide documents.

The name format:

It is suggested to follow a recent SIPB (Scientific Information Policy Board) recommendation and label all paper versions of CERN documents as CERN-(set)-yyyy-nnn. The convention for the relevant URLs is to keep the same name and replace the hyphens with slashes. To facilitate the path to this standardisation:

  • the URLs will be case insensitive
  • redirects will be in place to allow change of links/bookmarks.

The name components:.

URLs deriving from the above document names will naturally be:

http://cern.ch/ref/CERN/set/yyyy/nnn.

The 'CERN' (case insensitive) component can be omitted for simplicity.

Today's well-known URLs like http://consult.cern.ch/set (e.g. set=[writeup,cnl]) will remain valid for a 'generous' amount of time due to the above mentioned 'redirects'.

IT examples of "sets" are: IT/C5,IT/MCM, IT/GLM, group meetings and documents (IT/US), divisional projects (IT/L2K), IT Reports (IT). Till the registration of the "ref" site is completed some of the above "sets" exist as:

http://cern.ch/consult/CERN/IT/C5, http://cern.ch/consult/CERN/IT/MCM, http://cern.ch/consult/CERN/IT/reports

The nnn numbers will be automatically assigned by an easy-to-use web application.

Cross-boundary documents should carry no hierarchical (divisional) association; e.g. committees like FOCUS, Desktop Forum,ACCU etc could move for consistency to http://cern.ch/ref/CERN/committee/yyyy/nnn, e.g. http://cern.ch/ref/CERN/FOCUS/2000/008.

The files' location:

It is the author's choice whether the actual pages (files) are on afs or Windows, therefore transparently served by Apache (freeware web server running on Unix or NT) or IIS (Microsoft web server, NT). The submission/registration form allows reference to both filesystems and also to external URLs.

Advantages:

  • the URL is accepted in both upper and lowercase.
  • the naming allows us to easily offer index pages on every level.
  • the names are guessable.
  • the old names will still work due to redirects.
  • the URL is now an almost perfect match to its paper reference "CERN-IT-C5-YYYY-nnn" especially since the .html/.htm extentions will be dropped.
  • the "web upload" mechanism offered to the authors makes the web publishing system-independent.
  • as the name of any given document will be "nnn" under the "YYYY" directory, when a given document is composed of several parts, "YYYY/nnn/" will be a directory containing the document parts 'glued' together by Server Side Includes (SSIs) in a Welcome page.
  • authors will not have to change the physical locations where they edit their files today.
  • the automatic common mechanism to assign unique numbers will save today's manual steps by the secretariat.

Next steps:

Once the registration in "http://cern.ch/ref" is concluded, User Support could offer to "manage" the global "http://cern.ch/ref" site for 'sets' and 'committees' related to "User Documentation". The CERN top level pages (http://cern.ch/) are maintained by web.communications@cern.ch (ETT Division).

Minutes of meetings are considered by most people confidential and should not be searchable. Once the naming scheme is agreed, the search engine should be instructed (this is technically very simple if there is an unambiguous part in such URLs) to disallow indexing of such pages.

Some hints on authoring:

If templates for documents with standard layout, e.g. minutes of meetings, CNL articles, user guides etc. are judged useful they should be developed with the authoring tools recommeded at CERN. Common document "blocks" i.e. headers, footers, side bars are more easily maintained with the use of SSIs.

It is, of course, reminded that existing web documents, e.g. published articles or user guides shouldn't be edited online. The author should make a local copy and 'upload' only when satisfied with the result. Proper authoring tools like Dreamweaver or Frontpage are safe on this as the site is always built locally.