Is there a downside to multiple CSSs for modularity?
Barry Pearson
splitting the CSSs into smaller ones. So I may end up with (say) links to 2
CSSs for the same media type. (I'll be cascading them! I wondered how that
word crept in).
Is there a downside? For example, are there browsers that can handle my
current pages only because there is just 1 CSS, but would fail with 2 or more?
Are there any obstacles I need to be aware of?
I understand at least some of the performance issues, but I think there may be
a net gain. I'm not trying to do anything particularly novel. For example,
typically my home pages use less than half of the styles they current
download. And I have 8 CSSs for the photograph pages, differing only in a few
colour details, which is a development issue.
Thanks.
--
Barry Pearson
http://www.Barry.Pearson.name/photography/
http://www.BirdsAndAnimals.info/
http://www.ChildSupportAnalysis.co.uk/
Jim Ley
<ne...@childsupportanalysis.co.uk> wrote:
>All my pages now link to a CSS. For a couple of reasons, I'm considering
>splitting the CSSs into smaller ones. So I may end up with (say) links to 2
>CSSs for the same media type. (I'll be cascading them! I wondered how that
>word crept in).
I'd avoid it, it'll work, but it will make your site slower to load,
and slower to render.
If you want to author with seperate modules, have a publishing process
which simply piles all the files in together.
>I understand at least some of the performance issues, but I think there may be
>a net gain.
How so? The overhead of requesting the new page will almost certainly
swamp everything else, it'll prevent other things such as the image
being downloaded at the same time too.
Jim.
Barry Pearson
> On Tue, 23 Sep 2003 12:58:58 +0100, "Barry Pearson"
> <ne...@childsupportanalysis.co.uk> wrote:
>
>>All my pages now link to a CSS. For a couple of reasons, I'm
>>considering splitting the CSSs into smaller ones. So I may end up
>>with (say) links to 2 CSSs for the same media type. (I'll be
>>cascading them! I wondered how that word crept in).
>
> I'd avoid it, it'll work, but it will make your site slower to load,
> and slower to render.
It isn't quite as simple as that. Suppose I have the following (not
photographs in this case):
At the moment, a CSS of about 6 kilobytes used by the home page & 100s of
other pages.
With modularisation, the home page might use a 2 kilobyte CSS. The other 100s
of pages would have 2 CSSs, the 2 kilobyte CSS used by the home page, and a 4
kilobyte CSS that they would all share.
These would be cached in a typical browser. So the first access to the site,
typically to the home page, would be 4 kilobytes faster than now. The next
page looked at would have to get the extra 4 kilobytes, and have the extra
file access overheads. After that, both parts would be cached.
So, yes, if the viewer looks at 2 or more pages, there will be a slight loss
of peformance. But I keep reading just how important it is to grab the viewer
at the very first access. My reading is that the hardest part is to stop the
viewer leaving the site before s/he accesses the second page.
> If you want to author with seperate modules, have a publishing process
> which simply piles all the files in together.
Chuckle! I have - notepad & copy/paste! Where CSSs are concerned I'm using the
same tools the Victorians used.
>>I understand at least some of the performance issues, but I think
>>there may be a net gain.
>
> How so? The overhead of requesting the new page will almost certainly
> swamp everything else, it'll prevent other things such as the image
> being downloaded at the same time too.
Now think about the photograph pages. Each colour scheme (eg. dark grey
background, eggshell background) has a separate CSS at the moment. A few of
the styles in it set the colour (the background, the background image, border
colours, text colour). Most of them set other things that don't vary from one
photograph page to another. The first photograph would render slightly slower,
but any of a new colour would render faster.
The effect would actually be incredibly tiny compared with the size of the
photograph itself! My aim in this case would be to make it easier to maintain
the 8 separate CSSs I currently have for photographs:
http://www.birdsandanimals.info/web_site/formats.htm
As I said at the start of this thread, I have a couple of motives. One is to
improve home page loading time. The other is to reduce my own overheads.
I agree that it isn't easy. The raw performance figures will probably be as
you say.The user-satisfaction difference is harder to judge.
Jim Ley
<ne...@childsupportanalysis.co.uk> wrote:
>So, yes, if the viewer looks at 2 or more pages, there will be a slight loss
>of peformance. But I keep reading just how important it is to grab the viewer
>at the very first access. My reading is that the hardest part is to stop the
>viewer leaving the site before s/he accesses the second page.
downloading more than 1 file costs a lot, you have the extra bytes of
the request, and (potentially) the extra cost of making another
connection. There's also only 2 connections available so only 2 files
can be being downloaded simultaneously on even the fastest connection.
IE in a number of configs will almost always do if-modified requests
on even pretty cacheable files, so these cost on every page.
>The effect would actually be incredibly tiny compared with the size of the
>photograph itself! My aim in this case would be to make it easier to maintain
>the 8 separate CSSs I currently have for photographs:
>http://www.birdsandanimals.info/web_site/formats.htm
Stick them all in one file and turn on gzip handling, almost certainly
save more time-render-time than multiple files, and be easy as you're
only managing one file.
Messing about in 2k's remember the request itself from the browser is
over 1k, and can be even worse.
Jim.
Richard Watson
>> If you want to author with seperate modules, have a publishing process
>> which simply piles all the files in together.
>
> Chuckle! I have - notepad & copy/paste! Where CSSs are concerned I'm
> using the same tools the Victorians used.
Why?
There are loads of free editors out there.
How about chucking together stuff on the fly with your scripting
language of choice?
--
Richard Watson
http://www.opencolo.com/
High Quality, Value for money colocation
Liz
"Barry Pearson" <ne...@childsupportanalysis.co.uk> wrote:
>
> Is there a downside? For example, are there browsers that can handle my
> current pages only because there is just 1 CSS, but would fail with 2 or more?
> Are there any obstacles I need to be aware of?
>
Remember I'm your beta-tester on an old browser!
(Mind you, I'm a bit snowed under with the swen emails, but at least can now
get rid of a typical batch (30Mb must be my limit) in about 10 mins - unlike
at the weekend)
Just email me URLs when ready, and hope I don't delete you by mistake.
Liz
--
Virtual Liz at http://www.v-liz.co.uk
Kenya; Tanzania; India; Seychelles
New Aug '03: Namibia
"I speak of Africa and golden joys"
Barry Pearson
> On Tue, 23 Sep 2003 15:26:09 +0100, "Barry Pearson"
> <ne...@childsupportanalysis.co.uk> wrote:
>
>>So, yes, if the viewer looks at 2 or more pages, there will be a
>>slight loss of peformance. But I keep reading just how important it
>>is to grab the viewer at the very first access. My reading is that
>>the hardest part is to stop the viewer leaving the site before s/he
>>accesses the second page.
>
> downloading more than 1 file costs a lot, you have the extra bytes of
> the request, and (potentially) the extra cost of making another
> connection. There's also only 2 connections available so only 2 files
> can be being downloaded simultaneously on even the fastest connection.
> IE in a number of configs will almost always do if-modified requests
> on even pretty cacheable files, so these cost on every page.
OK. Pretty convincing for the 2 file case. But if I do stick to 1 file per
page, it may still be worth having a short version for the home page CSS of
one (only one) of my sites simply for the reason I identified above - the
impression gained by new viewers. (Once users get involved, life is never
easy, is it?!)
(Is it my imagination or was there a browser several years ago that made the
number of simultaneous fetches it could handle a configurable option, default
4?)
Thanks.
Barry Pearson
> "Barry Pearson" <ne...@childsupportanalysis.co.uk> writes:
[snip]
>> Chuckle! I have - notepad & copy/paste! Where CSSs are concerned I'm
>> using the same tools the Victorians used.
>
> Why?
>
> There are loads of free editors out there.
The fact that I started this thread shows that I don't yet know enough to know
what I want from a CSS editor. A month ago I didn't even know enough to have
asked the question! I actually already have 2 CSS editors, and feel that both
are worse than useless for my purposes. (Dreamweaver 4 has one, but I've
configured D4 to call notepad instead. And I have MS Frontpage2000 that came
with some other stuff and haven't got round to uninstalling yet).
Any suggestions?
> How about chucking together stuff on the fly with your scripting
> language of choice?
I think that reveals we have different IT backgrounds! I have & use scripting
languages within applications. (Dreamweaver recorded commands, Photoshop
actions, MS Office macros based on VB, etc). I don't use scripting outside of
specific application contexts.
However, you've prompted me to believe I have part of the answer in
Dreamweaver. My photograph page CSSs start with the rules where they differ
(typically colour), and the rest is a standard block identical in all of them.
So I can obviously (well, obvious now) use "find & replace" to replicate
changes I've made to any one of them.
Peter Connolly
> All my pages now link to a CSS. For a couple of reasons, I'm considering
> splitting the CSSs into smaller ones. So I may end up with (say) links to
2
> CSSs for the same media type. (I'll be cascading them! I wondered how that
> word crept in).
>
> Is there a downside? For example, are there browsers that can handle my
> current pages only because there is just 1 CSS, but would fail with 2 or
more?
> Are there any obstacles I need to be aware of?
>
> I understand at least some of the performance issues, but I think there
may be
> a net gain. I'm not trying to do anything particularly novel. For example,
> typically my home pages use less than half of the styles they current
> download. And I have 8 CSSs for the photograph pages, differing only in a
few
> colour details, which is a development issue.
>
Barry,
If it's good enough for Macromedia....
On http://www.macromedia.com, there is one CSS file included. If you look in
that CSS file, you'll see that it is a set of 8 imports, bringing in other
CSS files; for example, there's one for fonts, another for a global footer,
etc.; it looks like they change the base CSS file on each group of pages so
that it only imports the relevant CSS files.
This looks like Macromedia do exactly the same as you propose. I'd try it,
and see if it works.
HTH,
Pete.
--
---------------------------
Peter Connolly
Acute Computing
Derby
UK
Jim Ley
<ne...@childsupportanalysis.co.uk> wrote:
>OK. Pretty convincing for the 2 file case. But if I do stick to 1 file per
>page, it may still be worth having a short version for the home page CSS of
>one
Yep, that might be sensible.
>(Is it my imagination or was there a browser several years ago that made the
>number of simultaneous fetches it could handle a configurable option, default
>4?)
HTTP 1.1 says 2.
IE defaults to that, but it is modifiable.
IE uses 4 for HTTP 1.0 connections.
I don't know about other browsers.
Jim.
Barry Pearson
> In message <rCWbb.8053$Ne.32...@newsfep2-win.server.ntli.net>
> "Barry Pearson" <ne...@childsupportanalysis.co.uk> wrote:
> Remember I'm your beta-tester on an old browser!
In view of what Jim Ley says, I probably won't go for the 2 CSS approach.
However, I now have a page that I would like opinions on. (Hint!)
My photograph pages are pretty simply - a photograph, a description, and admin
at the bottom. They don't use tables for overall layout. (So I can claim to
have 100s of pages on the web that don't use tables for overall positioning!
Gosh). But they do use tables for each of the non-photograph bits. I'm trying
to decide whether to adopt a new template. Here is an example of it. It works
with IE6 and Mozilla Firefird with CSSs in action. It works - but I would like
opinions about how well it works - with Firebird & CSS off. I don't know
whether it works with other browsers, and would appreciate checks from anyone.
(Beware - the photograph is nearly 100 KB. It should get cached after the
first URL. I guess you've already seen it).
Current version:
http://www.barry.pearson.name/photography/pictures/k94/k94_14_25_3.htm
New version without any tables at all:
http://www.barry.pearson.name/test/k94_14_25_3_tableless.htm
CSS of new version:
http://www.barry.pearson.name/test/style_dark_grey_tableless.css
> (Mind you, I'm a bit snowed under with the swen emails, but at least
> can now get rid of a typical batch (30Mb must be my limit) in about
> 10 mins - unlike at the weekend)
I'm a bit concerned. I get perhaps 1000 a day, about half with the attachment
and about half that look like mail admin rejection notices (some purporting to
come from me). By now I was expecting that they would have died away, as
Sobig.F did after a few days.
> Just email me URLs when ready, and hope I don't delete you by mistake.
I'll respond here instead. Thanks.
Jim Ley
<noemail...@nospamrequired.com> wrote:
>On http://www.macromedia.com, there is one CSS file included. If you look in
>that CSS file, you'll see that it is a set of 8 imports, bringing in other
>CSS files; for example, there's one for fonts, another for a global footer,
>etc.; it looks like they change the base CSS file on each group of pages so
>that it only imports the relevant CSS files.
>
>This looks like Macromedia do exactly the same as you propose. I'd try it,
>and see if it works.
IT doesn't work for me, Macromedias pages are extremely slow to load
because of this behaviour.
They also use <script language="javascript1.2"> and that shows they
really don't know what they're doing, as that is absolutely the worst
choice to be made setting up some browsers to use non-compliant script
engines whilst others use compliant ones - it's not surprsing half
their script doesn't work.
Don't use macromedia generated content as best practice!
Jim.
Barry Pearson
> On Wed, 24 Sep 2003 08:58:38 +0100, "Peter Connolly"
>>This looks like Macromedia do exactly the same as you propose. I'd
>>try it, and see if it works.
> Don't use macromedia generated content as best practice!
I have to agree, although for a different reason. The Macromedia extensions
section now uses Flash for navigation, etc. My first problem is that I
normally run with Flash animation off, so have to reconfigure. (OK, not hard).
The second problem is simply that it is not what I am used to. I feel I have
had to learn something new, even though AFAIK they could have done it all with
bog-standard HTML. (I wonder how accessible it is to disabled people?)
Richard Watson
>> There are loads of free editors out there.
>
> The fact that I started this thread shows that I don't yet know
> enough to know what I want from a CSS editor.
I'm not sure you need a "CSS editor" but you do need a decent text
editor. At this point I usually recommend emacs, then some people try
it and most don't seem to like it, so I won't bother :-)
>> How about chucking together stuff on the fly with your scripting
>> language of choice?
>
> I think that reveals we have different IT backgrounds!
Probably not as different as you imagine, although going waaay back I
have some unix and VMS experience which you may or may not have. For
many years I was a slave to windows and its applications
> I have & use scripting languages within applications. (Dreamweaver
> recorded commands, Photoshop actions, MS Office macros based on VB,
> etc). I don't use scripting outside of specific application
> contexts.
Seems like about time you did then.
If you're prepared to give up about half an hour a day towards
learning some kind of scripting (perl,php,python,whatever) you'll find
that in about a week or two you'll know enough to save you time and
money later on. Instead of trying to find a shareware application to
perform those little tasks, you'll just knock up a script and have it
done and dusted in a few minutes.
Barry Pearson
>> enough to know what I want from a CSS editor.
>
> I'm not sure you need a "CSS editor" but you do need a decent text
> editor. At this point I usually recommend emacs, then some people try
> it and most don't seem to like it, so I won't bother :-)
If you are eating cornflakes, please stop now. I don't want you to choke on
them.
I already use a more powerful text editor for many purposes. It is called
"Word".
Chuckle!
[snip]
>> I think that reveals we have different IT backgrounds!
>
> Probably not as different as you imagine, although going waaay back I
> have some unix and VMS experience which you may or may not have. For
> many years I was a slave to windows and its applications
Before I became a business analyst, I did design & architecture of mainframes
& large scale computer systems. (Mostly VME from ICL).
I've used Windows for getting on for 15 years. I can't think of a good reason
to change. (I hope you hadn't restarted your cornflakes!)
[snip]
> If you're prepared to give up about half an hour a day towards
> learning some kind of scripting (perl,php,python,whatever) you'll find
> that in about a week or two you'll know enough to save you time and
> money later on. Instead of trying to find a shareware application to
> perform those little tasks, you'll just knock up a script and have it
> done and dusted in a few minutes.
I'll think about it, thanks. The trouble is that I don't know what those
little tasks are that I would stop using shareware for if I used a scripting
language. I think my needs are much simpler than yours, and I get by without
such shareware.
Richard Watson
>> I'm not sure you need a "CSS editor" but you do need a decent text
>> editor. At this point I usually recommend emacs, then some people try
>> it and most don't seem to like it, so I won't bother :-)
>
> If you are eating cornflakes, please stop now. I don't want you to choke on
> them.
</munch>
> I already use a more powerful text editor for many purposes. It is called
> "Word".
<splutter />
I don't even use a word processor for writing letters. Although that
might be because I don't seem to write many letters these days
compared to emails.
However a word processor is a wholly inappropriate thing to use for
editing plain text. If you don't already understand that, there's no
way I can explain it.
>> Probably not as different as you imagine, although going waaay back I
>> have some unix and VMS experience which you may or may not have. For
>> many years I was a slave to windows and its applications
>
> Before I became a business analyst, I did design & architecture of mainframes
> & large scale computer systems. (Mostly VME from ICL).
Aha. So you do know what a real computer looks like ;-)
> I've used Windows for getting on for 15 years. I can't think of a good reason
> to change. (I hope you hadn't restarted your cornflakes!)
I've got no problem with people using windows. However it does seem to
coexist with a particular view on life.
> [snip]
>> Instead of trying to find a shareware application to
>> perform those little tasks, you'll just knock up a script and have it
>> done and dusted in a few minutes.
>
> I'll think about it, thanks.
Do. Notwithstanding the fact that I'm sure you're already an
intelligent well-rounded person, it'll expand your horizons somewhat.
Try here:
http://honors.montana.edu/~jjc/easytut/easytut/easytut.html
and see how far you get ;-)
> The trouble is that I don't know what those
> little tasks are that I would stop using shareware for if I used a scripting
> language. I think my needs are much simpler than yours, and I get by without
> such shareware.
Well, maybe in your case it's expensiveware rather than shareware. You
did mention Dreamweaver, Frontpage and Word just now...
Seriously though, if you had any scripting experience you wouldn't
have even needed to ask your recent questions about modularising css
and html, you would have just written something to do it for you, be
it server-side or on your PC.
Barry Pearson
>> called "Word".
>
> <splutter />
>
> I don't even use a word processor for writing letters. Although that
> might be because I don't seem to write many letters these days
> compared to emails.
>
> However a word processor is a wholly inappropriate thing to use for
> editing plain text. If you don't already understand that, there's no
> way I can explain it.
You can't! The reason is it that I have been using Word (and before that,
WordStar 3 & 4 and then Word Perfect) for such purposes for many years (must
be nearly 2 decades) with a considerable amount of success! Even when I have
had a variety of text editors already on my PC. (As I probably have at the
moment).
Obviously I know some of the problems. A language-specific editor can check
and complete contructs. All real editors I've used manage such things as
indenting according to construct, etc. (And I do miss hex-editing!) But, for
example, when I've wanted to turn the sort of HTML output by Excel into simple
mark-up, I've found Word a great help, before pasting it into Dreamweaver.
[snip]
>> I've used Windows for getting on for 15 years. I can't think of a
>> good reason to change. (I hope you hadn't restarted your cornflakes!)
>
> I've got no problem with people using windows. However it does seem to
> coexist with a particular view on life.
Probably "go with the flow"? I tend to use industry-standard products or the
equivalent, such as Photoshop. Expensive, but I can get support, and the next
person I meet in the same line is likely to be using it too.
[snip]
> Do. Notwithstanding the fact that I'm sure you're already an
> intelligent well-rounded person, it'll expand your horizons somewhat.
>
> Try here:
> http://honors.montana.edu/~jjc/easytut/easytut/easytut.html
> and see how far you get ;-)
Chuckle! But can you write Fortran in Python? REAL programmers can write
Fortran in ANY language! (Nostalgia! I started on Egtran in about '66/67.
Rather like Fortran 4, with recursive subroutines, but simpler IF statements
like Fortran 2. I started using Assembler in '68, the day the Russians invaded
Czechoslovakia).
>> The trouble is that I don't know what those
>> little tasks are that I would stop using shareware for if I used a
>> scripting language. I think my needs are much simpler than yours,
>> and I get by without such shareware.
>
> Well, maybe in your case it's expensiveware rather than shareware. You
> did mention Dreamweaver, Frontpage and Word just now...
Indeed, that is the downside. (And I buy it in boxes too!) Probably the most
expensive was Photoshop. (I became a pariah on the Paint Shop Pro NG for
daring to suggest that while PSP was excellent value for money, it wasn't
actually as good as Photoshop for some serious work).
I think this arises from my mainframes background. There has always been a
very different mindset between small/medium systems people & mainframes
people. (Unfortunately, I was probably the one who went down the side street!
An Algol 68-like language instead of C/C++, OSI protocols instead of the IP
set, etc).
> Seriously though, if you had any scripting experience you wouldn't
> have even needed to ask your recent questions about modularising css
> and html, you would have just written something to do it for you, be
> it server-side or on your PC.
The question wasn't "how to modularise?" It was really "should I modularise?"
There is a trade-off for the viewer between performance on the home and
subsequent pages, as revealed by the discussion with Jim Ley. Until that is
resolved (probably in favour of sticking to one CSS per HTML page), the rest
is academic.
We are both right. But one of us takes the high road, and the other the low
road. (I don't know which is which).
Interesting discussion.
Barry Pearson
> Richard Watson wrote:
[snip]
>> Try here:
>> http://honors.montana.edu/~jjc/easytut/easytut/easytut.html
>> and see how far you get ;-)
>
> Chuckle! But can you write Fortran in Python? REAL programmers can
> write Fortran in ANY language!
I just found this:
http://mail.python.org/pipermail/python-list/2002-May/102434.html
And presumably most people know this:
http://www.pbm.com/~lindahl/real.programmers.html
Richard Watson
>> However a word processor is a wholly inappropriate thing to use for
>> editing plain text. If you don't already understand that, there's no
>> way I can explain it.
>
> You can't!
I feared as much :-/
> The reason is it that I have been using Word (and before that,
> WordStar 3 & 4 and then Word Perfect) for such purposes for many
> years (must be nearly 2 decades)
Well, I can't quite claim 2 decades. I was certainly using Word
Perfect in 1989 or so, about the same time that I learned to use
vi. Guess which one stood the test of time?
> Obviously I know some of the problems. A language-specific editor
> can check and complete contructs. All real editors I've used manage
> such things as indenting according to construct, etc. (And I do miss
> hex-editing!)
TBH that's not the real issue for me. In addition to all the above, I
need all these things in a text editor:
* Full operation without moving my hands from the keyboard.
* Full operation via remote virtual console login
* Lightweight enough that it takes no thought to launch it in an
instant for a quick edit
* Ability to test / compile programs without leaving the editor
* Complete configurability with regard to keystroke shortcuts etc
Handling all my mail/news is optional but convenient :-)
> But, for example, when I've wanted to turn the sort of
> HTML output by Excel into simple mark-up, I've found Word a great
> help, before pasting it into Dreamweaver.
This is where I'd probably write a script, or use a macro in emacs.
> I tend to use industry-standard products or the
> equivalent, such as Photoshop. Expensive, but I can get support, and the next
> person I meet in the same line is likely to be using it too.
I don't tend to view products as something which can be "industry
standard" or not. Products come and go, but real standards (like html)
develop and maintain interoperability between products.
>> http://honors.montana.edu/~jjc/easytut/easytut/easytut.html
>> and see how far you get ;-)
>
> Chuckle! But can you write Fortran in Python?
Indeed I would hope so, although there doesn't seem to be a goto
equivalent.
> REAL programmers can write Fortran in ANY language! (Nostalgia! I
> started on Egtran in about '66/67. Rather like Fortran 4, with
> recursive subroutines, but simpler IF statements like Fortran 2. I
> started using Assembler in '68, the day the Russians invaded
> Czechoslovakia).
My first FORTRAN was F77, but that was in about 1990 when F8X was long
overdue.
Anyway, if you can already hack fortran you should head straight for
here:
http://www.python.org/doc/current/tut/tut.html
> The question wasn't "how to modularise?" It was really "should I
> modularise?"
But of course you should. Although that doesn't mean you need multiple
stylesheets.
> There is a trade-off for the viewer between performance on the home and
> subsequent pages, as revealed by the discussion with Jim Ley. Until that is
> resolved (probably in favour of sticking to one CSS per HTML page), the rest
> is academic.
I think anyone who has to produce large amounts of web pages (or
anything really) realises eventualy that the "source" is not HTML and
CSS, but rather the source is used to create that.
Producing a single stylesheet per page would be a fairly costly
overhead too (an extra hit and download per page versus a stylesheet
already cached), but having 2-3 for the whole site in different
sections is not so bad. The question is how you arrive at them.
However if I was doing this I would consider including the styles in
the source of the HTML, thereby only adding a few hundred or so bytes
to each page's download rather than requiring the extra hit as
well. Then you could have your 1 stylesheet per page and make sure it
only has what you need in there for that page.
Having said that, I've tried a few different ways, and I think
limiting the number of styles used in the first place and throwing
them all in the same file is probably the most effective.
Barry Pearson
>> and subsequent pages, as revealed by the discussion with Jim Ley.
>> Until that is resolved (probably in favour of sticking to one CSS
>> per HTML page), the rest is academic.
> Producing a single stylesheet per page would be a fairly costly
> overhead too (an extra hit and download per page versus a stylesheet
> already cached), but having 2-3 for the whole site in different
> sections is not so bad. The question is how you arrive at them.
Whoops! My mistake.
I meant that any page would only use 1 CSS (of a particular media type). Along
with many other pages that would use the same CSS. Although ...
I have some places where an individual page has special styles on, because of
the topic being discussed. Eg. correspondence using a specialised text
styling. When I was eliminating "font" from my HTML over the last month, I
tried to think through what I was doing. But I was probably a bit too hasty in
putting those styles into the standard CSS used by many pages.
I'm debating whether I should have used special styles in the page itself, or
have an extra CSS just for that page and any later ones that need the same
style.
(Imagine a news website where all the articles use a standard CSS - as is
typical. Then one article needs a special style for certain words because of
content specific to that article. They wouldn't make the mistake of extending
the standard CSS. I wonder what they would do?)
Liz
"Barry Pearson" <ne...@childsupportanalysis.co.uk> wrote:
> Liz wrote:
> > In message <rCWbb.8053$Ne.32...@newsfep2-win.server.ntli.net>
> > "Barry Pearson" <ne...@childsupportanalysis.co.uk> wrote:
> [snip]
> > Remember I'm your beta-tester on an old browser!
>
> Current version:
> http://www.barry.pearson.name/photography/pictures/k94/k94_14_25_3.htm
>
> New version without any tables at all:
> http://www.barry.pearson.name/test/k94_14_25_3_tableless.htm
>
> CSS of new version:
> http://www.barry.pearson.name/test/style_dark_grey_tableless.css
>
I got 'no data can be found' on the top two, then my connection dropped.
I'll try again later.
..........
c1 hr later, still 'no data can be found'...
In fact, I can't even get onto your home page (I could before)
Is the problem at my end or yours?
Liz
"Barry Pearson" <ne...@childsupportanalysis.co.uk> wrote:
> I'm a bit concerned. I get perhaps 1000 a day, about half with the attachment
> and about half that look like mail admin rejection notices (some purporting to
> come from me). By now I was expecting that they would have died away, as
> Sobig.F did after a few days.
>
The techies on the csa groups seem to think it will run and run.....
Barry Pearson
> In message <9rccb.150$_N1.3...@newsfep1-win.server.ntli.net>
> "Barry Pearson" <ne...@childsupportanalysis.co.uk> wrote:
>> Current version:
>> http://www.barry.pearson.name/photography/pictures/k94/k94_14_25_3.htm
>>
>> New version without any tables at all:
>> http://www.barry.pearson.name/test/k94_14_25_3_tableless.htm
>>
>> CSS of new version:
>> http://www.barry.pearson.name/test/style_dark_grey_tableless.css
>>
> I got 'no data can be found' on the top two, then my connection
> dropped.
>
> I'll try again later.
>
> ..........
>
> c1 hr later, still 'no data can be found'...
> In fact, I can't even get onto your home page (I could before)
> Is the problem at my end or yours?
My end. I've just checked the hosting service & they say:
"Customers with email and websites hosted www2.fastweb.co.uk /
mail-relay-1.fastweb.co.uk will experience problems:
- sending / receiving email and with viewing websites and ftp
"Engineers are currently working on the problem, we currently have no
estimate."
Sorry! In 2 & half years I got a report once of a problem, and last week it
went slow for a day, but this is very serious.
Barry Pearson
> In message <9rccb.150$_N1.3...@newsfep1-win.server.ntli.net>
> "Barry Pearson" <ne...@childsupportanalysis.co.uk> wrote:
>> Current version:
>> http://www.barry.pearson.name/photography/pictures/k94/k94_14_25_3.htm
>>
>> New version without any tables at all:
>> http://www.barry.pearson.name/test/k94_14_25_3_tableless.htm
>>
>> CSS of new version:
>> http://www.barry.pearson.name/test/style_dark_grey_tableless.css
>>
> I got 'no data can be found' on the top two, then my connection
> dropped.
>
> I'll try again later.
It's back at last!
Aaaaaarrggh!
Liz
"Barry Pearson" <ne...@childsupportanalysis.co.uk> wrote:
>
> Sorry! In 2 & half years I got a report once of a problem, and last week it
> went slow for a day, but this is very serious.
>
No need to apologise!
Looking at my stats, there seem to be two days in every month when my visits
go down to almost zero (avarage 200 a day). I assume that means it's 'down'
but as it isn't regular, I haven't actually been able to catch them 'down'
yet!
Barry Pearson
> In uk.net.web.authoring on Wed, 24 Sep 2003 12:20:14 +0100, "Barry
> Pearson" <ne...@childsupportanalysis.co.uk> wrote:
>
> }Chuckle! But can you write Fortran in Python? REAL programmers can
> write }Fortran in ANY language!
>
> $i and $j used as indexes that did it.
>
> {R} FORTRAN didn't have a version number when I started with it.
Ha! I learned Fortran 4 in the pram, so I claim to be (a very young looking)
37. If you learned it in the womb, perhaps you are the same age.
When I first wrote the above, I had thought that the original article was
titled more or less what I said. Then I found that the article (below) was
actually called "Real Programmers Don't Use Pascal". This undermines my "real
programmer" status a bit, because I once used Pascal (MT+). In the early '80s
I wrote a laser printer driver for WordStar 3 so that I could produce
complicated architecture diagrams. It was my one and only use of Pascal,
honest! (Perhaps).
http://www.pbm.com/~lindahl/real.programmers.html
This rings a bell: "In theory, a Real Programmer could run his programs by
keying them into the front panel of the computer. Back in the days when
computers had front panels, this was actually done occasionally. Your typical
Real Programmer knew the entire bootstrap loader by memory in hex, and toggled
it in whenever it got destroyed by his program." Those were the days, with the
wonderful RCA 70/45 panel.
Owen Rees
<ne...@childsupportanalysis.co.uk> wrote in
<erFcb.192$%f1....@newsfep1-gui.server.ntli.net>:
>This rings a bell: "In theory, a Real Programmer could run his programs by
>keying them into the front panel of the computer. Back in the days when
>computers had front panels, this was actually done occasionally. Your typical
>Real Programmer knew the entire bootstrap loader by memory in hex, and toggled
>it in whenever it got destroyed by his program." Those were the days, with the
>wonderful RCA 70/45 panel.
I used to know the RIM loader for the PDP8 (octal rather than hex)...
starting at address 7756: 6032 6031 5357 - I can even remember what 5357
means!
--
Owen Rees - opinions expressed here are mine; for a full disclaimer
visit <http://homepages.tesco.net/~owen.rees/index.html#disclaimer>
for e-mail use "owen.rees at tesco.net" instead of the From address
Liz
Liz <l...@v-liz.co.uk> wrote:
>
> > Current version:
> > http://www.barry.pearson.name/photography/pictures/k94/k94_14_25_3.htm
> >
> > New version without any tables at all:
> > http://www.barry.pearson.name/test/k94_14_25_3_tableless.htm
> >
Onlyu difference on Fresco is that on the first version, the What, Where and
When question and answers are on the same line, but in the tableless
version, the answer jumps and looks like:
What?
Impala, (species neme in italics)
Where?
Kenya, Africa
When?
1994
(BTW, you wouldn't like to specify exactly where you saw each species, e.g.
which NP or whatever, would you?)
Barry Pearson
> In message <a5171a374c...@v-liz.co.uk>
> Liz <l...@v-liz.co.uk> wrote:
>>
>> > Current version:
>> > http://www.barry.pearson.name/photography/pictures/k94/k94_14_25_3.htm
>> >
>> > New version without any tables at all:
>> > http://www.barry.pearson.name/test/k94_14_25_3_tableless.htm
>> >
> Onlyu difference on Fresco is that on the first version, the What,
> Where and When question and answers are on the same line, but in the
> tableless version, the answer jumps and looks like:
>
> What?
> Impala, (species neme in italics)
> Where?
> Kenya, Africa
> When?
> 1994
Right! That is what I thought. Now which do you think is best?
In the original version (currently used for all my photographs), I simply use
a table for the description so that whether you use CSS or not you see it in
grid format.
In the potential new version, I use a definition list (which is logically a
bit flakey), then use the CSS to turn it into grid format like the table. So
if you don't use CSS, you get it as you say, which is a typical default for a
definition list. Also, I now accept that I should set italics via <i> </i>,
not via a style, because there isn't a logical choice, it really is italics.
So you see it as italics even without CSS.
And you also get the admin at the bottom in side-by-side grid format with the
original format. Whereas with the potential new version, I use CSS to float 2
separate <p>s onto the same line. So without CSS you should have seen 2
separate lines for those.
Preferences? (With CSS, they hardly look different from one-another).
> (BTW, you wouldn't like to specify exactly where you saw each
> species, e.g. which NP or whatever, would you?)
I'm trying to sort that out. I can't find my notebook! Aaaaaarrrggh! If I find
it, I will certainly update all the pages. It was a tour (I think arranged via
Thomas Cook) of a number of parks. One very good guide thoughout, Land Rover
Defender long wheelbase (laughing all the way at people in 2-wheel drive
vehicles). You mustn't economise on guide (if you have one) or vehicle.
James Taylor
Liz <l...@v-liz.co.uk> wrote:
>
> The techies on the csa groups seem to think it will run and run.....
Err, forgive my ignorance, but I can't work out what CSA stands for.
I don't think it's Child Support Agency, or Computer Security Announce,
or even CSS Sharing Area. A quick grep through my Usenet groups file
lists the following, and none of them would appear to be appropriate:
cmu.student.acf
comp.security.announce
comp.soft-sys.ace
comp.soft-sys.andrew
comp.soft-sys.app-builder.*
comp.software.arabic
comp.sources.acorn
comp.sources.amiga
comp.sources.apple2
comp.std.announce
comp.sys.acorn.*
comp.sys.alliant
comp.sys.amiga.*
comp.sys.amstrad.8bit
comp.sys.apollo
comp.sys.apple2.*
comp.sys.arm
comp.sys.atari.*
comp.sys.att
ctdl.sys.atari.st
ctdl.sys.atari8
cz.sci.alife
cz.sci.architekt
cz.soc.announce
Which "csa groups" are being referred to? Would someone enlighten me.
Thanks.
--
James Taylor, Cheltenham, Gloucestershire, UK. PGP key: 3FBE1BF9
To protect against spam, the address in the "From:" header is not valid.
In any case, you should reply to the group so that everyone can benefit.
If you must send me a private email, use james at oakseed demon co uk.
Liz
James Taylor <spam-block-@-SEE-MY-SIG.com> wrote:
> In article <13d71d374c...@v-liz.co.uk>,
> Liz <l...@v-liz.co.uk> wrote:
> >
> > The techies on the csa groups seem to think it will run and run.....
>
> Err, forgive my ignorance, but I can't work out what CSA stands for.
comp.sys.acorn (hierarchy).
(they apparently decided not to change the 'acorn' bit when Acorn went
AWOL.)
Liz
"Barry Pearson" <ne...@childsupportanalysis.co.uk> wrote:
> Liz wrote:
> > In message <a5171a374c...@v-liz.co.uk>
> > Liz <l...@v-liz.co.uk> wrote:
> >>
> >> > Current version:
> >> > http://www.barry.pearson.name/photography/pictures/k94/k94_14_25_3.htm
> >> >
> >> > New version without any tables at all:
> >> > http://www.barry.pearson.name/test/k94_14_25_3_tableless.htm
> >> >
> > Onlyu difference on Fresco is that on the first version, the What,
> > Where and When question and answers are on the same line, but in the
> > tableless version, the answer jumps and looks like:
> >
> > What?
> > Impala, (species neme in italics)
> > Where?
> > Kenya, Africa
> > When?
> > 1994
>
> Right! That is what I thought. Now which do you think is best?
Given a choice, I think the same line is better, but I wouldn't have thought
it was bad if I'd only seen the second version. *If* I'd noticed it at all
(which is unlikely unless I'd chosen it as one of the 'pages for analysis'
for the OU course I'm starting) I'd have wondered why the author hadn't put
them in a straight line, but it's hardly a hanging offence, and unlike most
CSS sites, it wouldn't have put me off.
> Also, I now accept that I should set italics via <i> </i>,
> not via a style, because there isn't a logical choice, it really is italics.
Absolutely.
Liz
jake
Pearson <ne...@childsupportanalysis.co.uk> writes
>
>This rings a bell: "In theory, a Real Programmer could run his programs by
>keying them into the front panel of the computer. Back in the days when
>computers had front panels, this was actually done occasionally. Your typical
>Real Programmer knew the entire bootstrap loader by memory in hex, and toggled
>it in whenever it got destroyed by his program." Those were the days, with the
>wonderful RCA 70/45 panel.
>
came with memory sizes so small that when your program was bootstrapped
into memory, the first instruction in your program -- and hence the
first instruction that it obeyed -- was a 'H' (Halt).
You could then use the toggle switches to manually enter the rest of
your program over the top of the bootstrap-loader.
Ahh .... the days when programmers were programmers -- and every memory
location counted.
And if you were to tell that to 'young folk' these days ...............
--
Jake
Barry Pearson
> In message <erFcb.192$%f1....@newsfep1-gui.server.ntli.net>, Barry
> Pearson <ne...@childsupportanalysis.co.uk> writes
> Ahh .... the days when programmers were programmers -- and every
> memory location counted.
>
> And if you were to tell that to 'young folk' these days
I still remember a meeting where we were discussing the design of the
archiving system for a mainframe operating system. We were deciding how to
store the date - how much disc storage to use for it.
We decided to hold just the last 2 digits of the year, because we were
convinced that people would not still be using the system on the 1st January
2000. But I lost track of just who was using the system. That is how easy it
was to build in a millennium bug!
The run up to the millennium was a bit nerve-racking! But if it caused any
problems, it appears to have killed all the witnesses at the same time. So
that is OK.
Alan J. Flavell
[...]
> We decided to hold just the last 2 digits of the year, because we were
> convinced that people would not still be using the system on the 1st January
> 2000. But I lost track of just who was using the system. That is how easy it
> was to build in a millennium bug!
I can report the opposite: in the 1960's I saw the operating system
code for dates in the Cambridge Titan, and there was a comment warning
that the code would fail in 2100. (No prizes for guessing why 2100
would be the first failure...)
I couldn't help thinking at the time how unlikely it was that the
particular code would be running even in 2000. Meantime, in 2000 I
had to set my much-younger video recorder to 1968 to get the right
date and day of the week (it's on virtual 1971 now of course).
all hopelessly OT for this group, sorry...
My "first" computer (met it in 1958) had separate paper tape readers
for each subroutine. Paste the tape into a loop, the reader chugged
once around the loop for each subroutine call, carrying out the
instructions as it went. "Tell the kids of today..." ;-)
Isofarro
> I can report the opposite: in the 1960's I saw the operating system
> code for dates in the Cambridge Titan, and there was a comment warning
> that the code would fail in 2100. (No prizes for guessing why 2100
> would be the first failure...)
Hmmm, no obvious guess... well Unix timestamps (counting number of seconds
from 1 January 1970) overflows its long integer definition somewhere in
2038 (IIRC), so it doesn't sound like that particular problem.
Sounds like a byte field being used as a year storage
in an offset format,
where 1960 is 0 and 2100 is 140. Using a signed byte gives 128 year lifespan
before the inevitable time-travel (0-127). Using an unsigned byte offers 256
years which takes it at least to the year 2215. Hmmm... that doesn't fit
nicely into your puzzle above.
I'm curious now!
--
Iso.
FAQs: http://html-faq.com http://alt-html.org http://allmyfaqs.com/
Recommended Hosting: http://www.affordablehost.com/
Web Design Tutorial: http://www.sitepoint.com/article/1010
Jim Ley
<spam...@spamdetector.co.uk> wrote:
>Alan J. Flavell wrote:
>
>> I can report the opposite: in the 1960's I saw the operating system
>> code for dates in the Cambridge Titan, and there was a comment warning
>> that the code would fail in 2100. (No prizes for guessing why 2100
>> would be the first failure...)
>
>I'm curious now!
some time around the 60th day would be my guess...
Jim.
Alan J. Flavell
> Alan J. Flavell wrote:
>
> > I can report the opposite: in the 1960's I saw the operating system
> > code for dates in the Cambridge Titan, and there was a comment warning
> > that the code would fail in 2100. (No prizes for guessing why 2100
> > would be the first failure...)
>
> I'm curious now!
It had an over-simplified rule for calculating leap years (it ignored
the 100-year and 400-year rules), so it got the right answer for 2000
for the wrong reason, but failed for 2100 (and for 1900 also, but
since that particular code was only meant to deal with the current
date, that didn't matter).
cheers
Dr John Stockton
news:uk.net.web.authoring, Isofarro <spam...@spamdetector.co.uk>
posted at Sat, 27 Sep 2003 12:13:21 :-
>Alan J. Flavell wrote:
>
>> I can report the opposite: in the 1960's I saw the operating system
>> code for dates in the Cambridge Titan, and there was a comment warning
>> that the code would fail in 2100. (No prizes for guessing why 2100
>> would be the first failure...)
That's like, IIRC, Caldera DOS, which however has a comment bug for that
day.
>Hmmm, no obvious guess... well Unix timestamps (counting number of seconds
>from 1 January 1970) overflows its long integer definition somewhere in
>2038 (IIRC), so it doesn't sound like that particular problem.
2038-01-19 03:14:08 GMT, IIRC. Confirmed, 2^31 s from start 1970. Tue.
2^32 s is 2106-02-07 Sun 06:28:16 GMT.
>Sounds like a byte field being used as a year storage
See via below, including dayscale and critdate.
Any opinions on "AD Integer8" time?
--
© John Stockton, Surrey, UK. ?@merlyn.demon.co.uk Turnpike v4.00 MIME. ©
Web <URL:http://www.merlyn.demon.co.uk/> - w. FAQish topics, links, acronyms
PAS EXE etc : <URL:http://www.merlyn.demon.co.uk/programs/> - see 00index.htm
Dates - miscdate.htm moredate.htm js-dates.htm pas-time.htm critdate.htm etc.
Alan J. Flavell
Hi John, I had no doubt that this thread would "pull your string", and
I wasn't wrong ;-)
> JRS: In article <2tu3lb...@sidious.isolani.co.uk>, seen in
> news:uk.net.web.authoring, Isofarro <spam...@spamdetector.co.uk>
> posted at Sat, 27 Sep 2003 12:13:21 :-
> >Alan J. Flavell wrote:
> >
> >> I can report the opposite: in the 1960's I saw the operating system
> >> code for dates in the Cambridge Titan, and there was a comment warning
> >> that the code would fail in 2100. (No prizes for guessing why 2100
> >> would be the first failure...)
[...]
> >Hmmm, no obvious guess... well Unix timestamps (counting number of seconds
> >from 1 January 1970) overflows its long integer definition somewhere in
> >2038 (IIRC), so it doesn't sound like that particular problem.
>
> 2038-01-19 03:14:08 GMT, IIRC. Confirmed, 2^31 s from start 1970. Tue.
> 2^32 s is 2106-02-07 Sun 06:28:16 GMT.
>
> >Sounds like a byte field being used as a year storage
Yes, but that's far away from the Cambridge Titan, which had 48-bit
words[1], and the problem (as noted already on this thread) was about
calendar dates, rather than epoch seconds.
[1] if memory serves me[2] (the paperwork is too deeply buried), it
had 24-bit character addresses, made up of a 21-bit word address and a
3-bit "octal fraction" for the character address.
[2] As usual, Google also knows:
http://www.google.com/search?q=titan+prototype.atlas.2
Liz
"Barry Pearson" <ne...@childsupportanalysis.co.uk> wrote:
> Thanks for that. Since you prefer those 100s of pages across 2 web sites
> the way they are, I have decided to let you twist my arm into leaving them
> the way they are!
Just send the cheque to 'Madame Armtwist'.
> Ouch! And there I was, desperate to waste lots of time changing them, before
> you persuaded me otherwise.
So now you've got to make good use of the saved time.....
Barry Pearson
> In message <6Q2eb.778$iH6...@newsfep1-gui.server.ntli.net>
> "Barry Pearson" <ne...@childsupportanalysis.co.uk> wrote:
>
>> Thanks for that. Since you prefer those 100s of pages across 2 web
>> sites the way they are, I have decided to let you twist my arm into
>> leaving them the way they are!
>
> Just send the cheque to 'Madame Armtwist'.
Er ... I get spam that says things like that! Is it you?
>> Ouch! And there I was, desperate to waste lots of time changing
>> them, before you persuaded me otherwise.
>
> So now you've got to make good use of the saved time.....
"I saved lots of money at the sale today".
"Show to me".
Liz
"Barry Pearson" <ne...@childsupportanalysis.co.uk> wrote:
> Liz wrote:
> > In message <6Q2eb.778$iH6...@newsfep1-gui.server.ntli.net>
> > "Barry Pearson" <ne...@childsupportanalysis.co.uk> wrote:
> >
> >> Thanks for that. Since you prefer those 100s of pages across 2 web
> >> sites the way they are, I have decided to let you twist my arm into
> >> leaving them the way they are!
> >
> > Just send the cheque to 'Madame Armtwist'.
> Er ... I get spam that says things like that! Is it you?
hahaha not me.
I'm still too busy deleting the stuff.
> >> Ouch! And there I was, desperate to waste lots of time changing
> >> them, before you persuaded me otherwise.
> > So now you've got to make good use of the saved time.....
> "I saved lots of money at the sale today".
> "Show to me".
:-)))