Unfortunately I cannot attend the WWW workshop in the US. I am
looking after the needs of users (mainly in the CERN environment, but
believe they are not special). I have here summed up the most urgent
wishes for development of WWW in the near future. I would like you to
consider them at the meeting, and wish you all a fruitful gathering,
brainstorming and wise reflection.
Robert.
Among the wishes, two stand out far more strongly than any others:
1) how do authors who are not computer literate put information
into the web?
For this, we need:
- a portable HTML editing object plus some ancillary objects for
setting styles etc.
The moment we have html editing, a large number of people will be
able to provide urgent information themselves as if they were using a
word processor. A number of features now proposed or in use will also
become unified and/or irrelevant: hotlists, document lists, saving
and organising of URLs and all these schemes dreamed up by those who
could not from the first use Tim's browser for the NeXT, and have not
experienced direct WWW authorship.
I would consider any transformation of files in existing formats as
useful but not adequate in the longer term.
- simpler server installation and configuration.
- serving of different sets of documents from the same machine on the
same port 80.
Today it is difficult to organise both html directories and several
indexes from the same machine unless you use different ports and run
many servers.
I also take the opportunity to repeat that there is a grant available
for any interested developer to work at CERN (provided he/she holds a
passport of a CERN member state).
2) how do we protect the documents that belong only to our workgroup
from unauthorised access?
This we are doing now (Ari Luotonen, luotonen@dxcern.cern.ch)
according to the design of http2, and thus it is being taken care of.
Then there are to my mind a number of smaller items but which are
important for the continued future:
- anything that will unify and simplify the URLs.
Do not let them grow into large unintelligible things. When I see a
URL, I want to have an idea of what it is that I'm going to get.
Is there really a need to have so many protocol specifiers? Can some
be unified with others?
- indirect documents (to deal with both relocations and references,
most of that is defined in the http protocol now).
This must also include program documents: the URL specifies the
script to be run by the server at the server site, e.g. complicated
Oracle queries.
- decide on a (very) few formats for graphical material, and
implement their viewers as part of the browsers. This does not
exclude that any arbitrary format can be handled within a closed
community through the general format negotiation.
However, I do not wish to accumulate tens of viewers just because
there are pictures out there in all those formats. Providers also do
not wish to waste storage to keep copies in multiple formats.
A few wisely chosen formats displayed inside the browser window (i.e.
without launching separate processes) would certainly drive the
provider community in a direction.
Personally, I see two main graphics uses: (1) drawings (scalable and
composed of editable objects, like technical drawings and diagrams)
and (2) images (bitmaps, like photos).
Having seen some of the uses, I think for drawings PostScript (and
the new PDF format of Adobe) are essential. While some packages are
notoriously bad for generating huge PostScript files, good drawing
packages such as Illustrator and xfig generate very compact and
well-behaved files.
For images clear contenders are GIF and JPEG.
--------------------------
Robert Cailliau cailliau@cernnext.cern.ch
World-Wide Web Project
CERN -- European Laboratory for Particle Physics
CH-1211 Geneve 23 (Switzerland)
Tel. +41 22 767 5005