HP3000-L Archives

February 1997, Week 1

HP3000-L@RAVEN.UTC.EDU

Options: Use Monospaced Font
Show Text Part by Default
Show All Mail Headers

Message: [<< First] [< Prev] [Next >] [Last >>]
Topic: [<< First] [< Prev] [Next >] [Last >>]
Author: [<< First] [< Prev] [Next >] [Last >>]

Print Reply
Subject:
From:
Neil Harvey <[log in to unmask]>
Reply To:
Neil Harvey <[log in to unmask]>
Date:
Tue, 4 Feb 1997 10:44:28 +0200
Content-Type:
text/plain
Parts/Attachments:
text/plain (178 lines)
Wirt mentions some good stuff re web development............

Best advice (IMHO) is to test your pages on more than one browser. We
developed a site for Bankmed http://bankmed.co.za using IIS on NT and MS
IE3.1. All looked great, until I ran it through Netscape. The tables
were all over the place, and had to be reworked extensively.

Also, my advice is NOT to use any of the user friendly HTML page
generators for anything other generating sample code, which you can cut
and paste into your beautifully hand-crafted pages.

I say this because I have seen quadrupling of the size of some pages
simply by running it through one of the well known editors, and this
"bloat" must surely be contributing to the great internet slowdown.

We are busy rolling out Quiz reports onto an Intranet running - wait for
it - Apache serevr on an HP3000. It works really well.
We have a few "include" files and with a minimum of changes, we can get
a report that traditionally printed on paper, converted into a very
acceptable web report.

We also have some command files that "publish" the report into a special
dircetory, and "publink" the report to users, so that they simply click
on a link on their win95 desktops to get to their own reports. We use
one Cobol program to generate nice looking html pages that index the
various directories, because the default web browsers view of a
subdirectory is VERY unfriendly, and thi is run every time a new report
is "published" or "publinked"

So the Quiz code looks like this:-

        access somefile

        use inithead.qz                                 ;start up HTML
stuff
        use pagehead.qz                         ;some standard headings

        report all                                      ; and other
stuff

        use pagefoot.qz                                 ;stuff at the
end of each page, if required
        use finfoot.qz                                  ;close off the
page

        set report device disc name REPORT1
        set rep limit 100000
        set page width 132 length 0
        go

        publish REPORT1 /web/Reports/Daily/ Report1
        publink nharvey
        publink jvonbulow
        publink mduplessis
        publink hrhelizabeth

The report is "published" once, and ever user has a link to it.


What I'd like, of course, is to be able to say to Quiz:-

        set report device HTML name
/web/Reports/Nicely_Formatted_Quiz_Report.htm

But I guess that'll all come later. :)

Regards

Neil


> -----Original Message-----
> From: HP-3000 Systems Discussion [SMTP:[log in to unmask]] On
> Behalf Of Wirt Atmar
> Sent: Tuesday, February 04, 1997 9:35 AM
> To:   [log in to unmask]
> Subject:      Re: web site development
>
> In order to complete what I said earlier about web page development,
> let me
> add what I had intended to say in my first posting. There are three
> items
> that I wish to somewhat modify:
>
> The first is that when lifting the HTML code from a remote web site,
> you
> don't have to "scrape" the entire screen with the I-bar. Rather, use
> Control-A to "select all" before you Control-C ("copy selected text to
> clipboard"). You only need the I-bar scraper if you are lifting a part
> of the
> code. I mention all of this because these editing features are hidden
> (or at
> least not explained) in the Netscape source code window. (BTW, the
> Control-
> type commands are borrowed directly from the Macintosh -- and are
> ancient
> there -- but they are now beginning to become almost as uniform on new
> PC-based software as they were on the Mac. Almost all of Microsoft's
> newer
> products use these sequences, and that actually helps bring the same
> uniformity of use that was always characteristic of the Macintosh and
> made it
> such a pleasure to use).
>
> The second item is my recommendation for looking at the source for our
> web
> page. I mention that because you can see reasonably elaborate pages
> that are
> completely hand-assembled. That hand-assembly aspect will be
> particularly
> clear once you view the source code because of the non-uniform lengths
> of the
> individual lines of text. As the content of the text is modified
> through
> revision, the lines are made shorter or longer as necessary. But it
> doesn't
> matter to HTML. And editing the text content of a web page in this
> manner is
> no more difficult than creating and editing an e-mail message.
>
> [A note about HTML: a line break (a carriage return) is handled as a
> space in
> HTML -- and multiple spaces, should they be found in the text -- are
> automatically reduced to a single space, thus the various line lengths
> are
> irrelevant to HTML. If you wish multiple spaces to appear in your
> text, you
> must use the sequence, &#160, a non-reducible space (information which
> will
> appear in the Beginner's Guide to HTML that I earlier referenced). To
> end/begin a new paragraph, you use the <p> tag. To cause a line to
> break, you
> use <br>. Capitalization within the tags is generally irrelevant to
> HTML.]
>
> After you've looked at the source for our web page, go look at as many
> others
> as you care to, especially the ones that you find attractive. It will
> be easy
> to see that most pages are machine-generated -- simply because of the
> uniform
> line lengths and the appearance of some semi-automated annonations
> that
> appear throughout the code. But all of this doesn't make difference to
> the
> HTML interpreters that appear in the various browsers. The web pages,
> hand-assembled or automatically assembled, will look the same on the
> CRT's
> screen if your code is the same.
>
> The third item that I wish to elaborate on is that you can't trust a
> web page
> that is only viewed during its development with one browser. I now
> believe
> that it is imperative to gather up at least three browsers: Netscape,
> IE, and
> AOL to see how they each individually render your web page. There will
> be
> differences -- some of which may be occasionally very large indeed.
> But, with
> experimentation, you can generate code that is acceptable to all three
> and
> which will be rendered essentially identically.
>
> And, finally, you must decide what minimum level of HTML you are going
> to
> support. We have standardized on HTML 2.0, which is now reliably
> supported by
> Netscape 3.0, IE 3.0, and AOL 3.0. We decided early on -- as has
> almost every
> other commercial web site -- not to bother with text-only browsers
> such as
> Lynx.
>
> Again, best of luck,
>
> Wirt Atmar

ATOM RSS1 RSS2