|
The new BaBar web
Introduction
On the 10th of February 1999, after a period of testing, the new BaBar
web was finally released. This web page contains instructions on:
Features of the new web are also outlined, along
with a note on the new web's future.
If you have any questions, or you find any problems, then please contact the WebMaster via e-mail or use our hypernews forum.
The WebTeam would like to thank all those who provided feedback on the trial version of the web.
The BaBar web banner and navigation sidebar (a.k.a. the "wrapper" or "header")
Most BaBar web pages contain a banner that runs across the top of the page
and a navigation sidebar that runs down the left-hand side of the page.
The top banner's purpose is to make it easier to navigate
between the main branches of the BaBar web (Physics, Detector,
Computing, Organization, etc.), to
switch between the RAL mirror site and the SLAC site,
and to provide access to a few other commonly-used
links. The purpose of the left-hand sidebar is to
provide easier navigation within a particular branch of the web. The
banner is the same on all pages, but the content of the sidebar
varies, depending upon which major branch the page resides in,
and possibly where
in that branch.
The process of adding the banner and navigation sidebar to
a page is called "wrapping". See below
for details on how to add, remove, or
update the wrapper on a single page (but read the discussion in the next
paragraph regarding the automatic wrapping script that runs every night).
Automated Wrapping Script
The banner and navigation sidebar are automatically updated on the SLAC site
from master copies each night.
The next synchronization of the RAL mirror will then migrate the
updates to RAL. A trscrontab job at SLAC, running under userid "rfc", performs
the automatic update, starting at 21:00 local (Pacific) time. The automatic
update performs the following:
- Adds the banner/sidebar to any new files that have appeared.
- Replaces the banner/sidebar on any existing pages that have been edited
since the last time the wrapper script inspected the page.
- Updates the banner/sidebar on each page whenever the corresponding
master copy of the header has been modified.
- Preserves timestamps of files during the wrapping process
(otherwise date-sorted search engine results would be meaningless, as all
pages would have the date of the latest change to the banner, not the date of
the latest change to the content).
Changes to the master header files should be coordinated
with the BaBar webmaster, Ray Cowan, or with Charlotte Hee.
Note in particular
that changes made to the banner/sidebar text as it appears in a particular
webpage's file will be lost when the
automatic update script runs that evening, as it will re-insert the header
from the master copy.
Changes should be coordinated
with the webmaster to make sure they are applied to the master copies.
Broken Web Pages and the Wrapping Script
Some pages are not wrapped, or may contain very outdated versions of the header,
because of HTML syntax errors in the file. The wrapper script has to be able
to reliably identify the head and body content
of the page. If the page is missing
certain basic HTML tags, or they are incorrectly placed, the wrapper script
will abort and leave the page unchanged. When you edit BaBar web pages,
be careful to follow proper HTML syntax. The most common ways pages become
broken are:
- Using an existing, broken page as a template for a new page. This propagates
the problem to a new page.
- Using Microsoft Office (MSO) products to edit web pages. Although our
header HTML code passes validation by the W3C validator, some versions of
Microsoft Office products, such as Word and FrontPage, will still corrupt
the header. They frequently will
insert large numbers
unneeded HTML tags that provide support for some MSO features
(e.g., "<o:p>" and "<mso:...>"). Sometimes they
introduce other errors in the existing header code (such as
uncommenting HTML comments, or commenting out portions of active HTML
code) that break the page
and make it unwrappable by anything other than human hands.
To prevent problems with MSO products,
it is recommended that before editing any BaBar web
page using a Microsoft product, the web page be
unwrapped before
editing, and re-wrapped when returned to the BaBar web.
(When returning the edited page to the web, you could also just
leave the page unwrapped and let the automated script take care of it the next
time it runs.)
To unwrap the page, issue the command "WEBwrap -rh mypage.html" from a Babar unix account. To restore the header, use the same command, but specify
the option "-ah" instead of "-rh". These represent "add
header" and "remove header", respectively.
The webmaster will be happy to
provide a list of broken pages if you are concerned that a page you are
interested in is
so affected. You can check a specific page yourself to see if it is
broken by running the
W3C Validator
(linked at the bottom of the left-hand sidebar) or by running
the testhtml program
from a BaBar unix account, giving the filespec of the webpage as an
argument (e.g., "testhtml -e mypage.html". Warnings are ok, but if
it reports errors, the page is broken).
Some pages in certain areas of the BaBar web are also not wrapped by prior
arrangement with the webmaster. These include pages that are generated
automatically by other scripts for which the wrapping process may cause
problems. Other exceptions are in a few specific areas of the web that have a
different appearance for a specific reason, or in certain
other cases. However, most pages are wrapped to enhance navigation,
provide a similar
look-and-feel to the whole of the BaBar web, and hold to a minimum
the number of sub-webs springing up of varying designs.
-
If you use Internet Explorer then you will first of all need to convert your favourites file to Netscape-type bookmarks file using the free favourites-to-bookmarks two-way
converter available from Microsoft.
-
Copy your bookmarks to a backup file (e.g. oldbookmarks.html) and use the bookmarks-changing tool on your bookmarks.html file - i.e. type:
$BFROOT/www/Computing/WebTeamPages/tools/bin/FixBookmarks bookmarks.html
-
Check to see whether your new bookmarks file works by closing down Netscape, copying your converted bookmarks.html to your .netscape directory and clicking around.
-
Internet explorers can convert their new bookmarks file back to favourites if they wish to.
- If you are experienced in writing HTML then create a new page by saving and editing the template web page, or, if you are inexperienced or unsure, use the example web page
which contains detailed instructions.
- Copy your new page to a suitable directory under $BFROOT/www/ on the SLAC server (which contains a header.inc file) and wrap it
with BaBar navigation bars using the page-wrapping tool. E.g.
WEBwrap -ah mypage.html
 To wrap all the files starting from a root directory ,<ROOTDIR>, type:
WEBwrap -ah -d <ROOTDIR>
- Test the validity of your HTML using our testhtml tool. E.g.
$BFROOT/www/Computing/WebTeamPages/tools/bin/testhtml -e mypage.html
Another more comprehensive way to check your HTML is to use the HTML
validator at
validator.w3.org.
selecting "HTML Transitional 4.01" as the
document type, and "iso-8859-1"
as the character encoding. You can add these to the page directly, by placing
<!DOCTYPE html PUBLIC "-//W3C//DTD HTML 4.0 Transitional//EN">
at the top of the file, before the "<html>" tag, and
<meta http-equiv="Content-Type" content="text/html; charset=iso-8859-1">
between the <head>...</head> tags.
These tags are included
in the example web page mentioned above.
- Test the links in your web page,
list of pages, or pages within a directory and its subdirectories
using our CheckLinks tool. E.g.
To check the links in mypage.html, type:
CheckLinks mypage.html
To check the links in all pages within a directory <ROOTDIR> type:
CheckLinks <ROOTDIR>*.htm*
To check all pages within and below a directory <ROOTDIR> type:
CheckLinks -d <ROOTDIR>
- Review the guidelines and tips for present and future web authors.
From your browser you can load in a page and click the "View printable page"
button in the navigation bar, which uses a CGI script to remove the
wrapper from a page so you can print just its content. Alternatively, you can
print a page called "mypage.html"
without the navigation bars by typing:
$BFROOT/www/Computing/WebTeamPages/tools/bin/WEBprint mypage.html
N.B. Images on the page are not printed with this version of the tool.
Web page protection is done on a directory basis wherein a .htaccess file
is required. Until recently, protected web pages used a .htaccess file which
restricted access to users such as babar, techbd, and execbd which have
well-known passwords. This method has been discontinued.
Now use the HyperNews method to protect a directory as follows:
To get Hypernews password protection:
cd <Directory to be protected>
ln -s /afs/slac.stanford.edu/g/babar/www/internal/.htaccess .htaccess
To make directory visible only from SLAC machines but without the password protection:
cd <Directory to be protected>
ln -s /afs/slac.stanford.edu/g/babar/www/internal/.htaccess.SLAC .htaccess
IMPORTANT NOTE: in both the above examples, you must use
/afs/slac.stanford.edu/... and NOT /afs/slac/... Otherwise the
RAL mirror site directories will not be protected!
Note that this only protects the files from unauthorized web access.
All files in the directory with the .htaccess file and all its
subdirectories should also be protected from normal file system access by
AFS users. The AFS command 'fs sa' can be used to protect each of the
directories. Typically, from the directory where you placed the
.htaccess file, you will need to run:
find . -type d -exec fs sa {} system:slac l system:authuser l \;
This gives general SLAC AFS users list-only access. If you do not
want others to see the file names, replace "l" with "none" above.
Unfortunately, the above procedure denies access to the web servers.
To re-enable access to the web servers run:
find . -type d -exec fs sa {} g-babar:host-www rl g-www:hosts-servers rl \;
Please note this critical difference between web and AFS directory
protection. The .htaccess file for web security protects the directory
it is in and all subdirectories beneath it. Changes to the .htaccess
file affect all these directories. In AFS, changes to the ACLs of a
parent directory do not affect the ACLs of its subdirectories. New
subdirectories do inherit the ACLs of its parent though.
If you want some text in your BaBar web page to appear on its own page when printed
then put the text inbetween the following pairs of html tags as follows:
<!-- WIT page size table start one -->
<!-- WIT page size table start two -->
TEXT PUT HERE WILL APPEAR ON ITS OWN PAGE WHEN PRINTED
<!-- WIT page size table end one -->
<!-- WIT page size table end two -->
If you load the resultant web page, wrapped with standard navigation bars, and then click
on the "View Printable page" button, you get a page with no navigation
bars, of course, but also the text you surrounded with the tags is placed on its own page
when printed.
The new BaBar web has the following features which should make it easier to navigate around, and to find the information you require:
- A streamlined directory structure under $BFROOT/www/.
- No (or very few) broken links. (With your help we have fixed many broken links so they point to the intended URLs, but, certain formerly broken links now point to a token page in the new web, because the intended
URLs are still unknown.)
- No symbolic links within the new file structure.
- An archive web area under $BFROOT/old-www/ with the same file structure as $BFROOT/www/.
- Navigation bars on every page from which the main BaBar areas and topics relating to the page you are on are just a click away.
- A much improved search engine available from every page with the ability to search only that part of the Babar web space you are interested in.
- A sitemap for the web available from every page.
- A subscription form on every page so you can receive e-mail when the pages you frequent are moved or updated.
- A link on each page to both the SLAC original and the mirror image of the page at RAL.
- No frames: all frames have been replaced have been replaced by tailor-made header files which contain the same information as the frames. Furthermore, all links to the frames have been replaced. The various reasons for this policy are
described in the authorship guidelines.
The web improvement is an ongoing process - carried
out by the Webmaster, but requiring a great deal of cooperation from BaBar web authors - with the aim of
providing a well organized web in which information is up-to-date, easy to read and easy to find.
|
