Forum Moderators: open

Message Too Old, No Replies

quoteless html

         

Natko

12:16 am on Mar 28, 2004 (gmt 0)

10+ Year Member



Hi,
some says that are mandatory, some says optional.
I have my html pages quoteless in function of optimization and works fine in IE5.5. Unfortunately, I haven't tested them on other interpreters, yet.
other expirience in this?
:-)

ergophobe

12:01 am on Apr 1, 2004 (gmt 0)

WebmasterWorld Senior Member 10+ Year Member Top Contributors Of The Month




I feel like I'm the only currious webmaster moron?

I think most of us are curious and many of us are... Well, I should only speak for myself (and that by the way is irony, not sarcasm. Important difference).

No, I think why this thread has attracted so much attention is because many of us are very curious. What we are curious about, however, is why you would go to great pains to reject standards for such a small payoff.

I just checked a pretty large page of mine and the savings was 168 bytes.

Natko

3:42 pm on Apr 1, 2004 (gmt 0)

10+ Year Member



That is not so much, I agree. But do you need them?
If, yes, why?
Becouse, bald tags are ugly to you, all of your audience are DSL, Inktomi have dificulties with some bal '/', NS3.x are confused with bald hrefs, or will missdisplay some non safe bald color tags?
Even if I think I'll requote all of my #, /, and hrefs,
we are on 4th page now, and I haven't heard any proof or experience (maybe one) yet, only a w3c recomms.
Let me wait until next index.

grahamstewart

4:32 pm on Apr 1, 2004 (gmt 0)

WebmasterWorld Senior Member 10+ Year Member



Why bother with any of the w3c recommendations then?

This page works perfectly in IE6 and Opera and I'm sure Google will be able to read it.


<keywords=widgets,grommits,sale,retail,shop,stuff>
<language=Queens English old chap>

<h1>Welcome to widgets.com</h1>
We sell:
<menu>
<a href=red.html>red widgets,
<a href=blue.html>blue widgets,
<a href=green.html>green widgets.</a>

<title>Widgets for Sale - widgets.com</title>

<center><small><plaintext>{site designed by CuriousMoron Ltd}

;)

Natko

4:50 pm on Apr 1, 2004 (gmt 0)

10+ Year Member



:-))))

R1chard

6:35 pm on Apr 1, 2004 (gmt 0)

10+ Year Member



It's not just a color that might be misinterpreted. Sometimes even if there's just a slight error with some italics, then you get a blank page.

You want browser examples of links that don 't work without quotes?

Nothing worked in Amaya 6.1
# didn't work in OffByOne 3.4 (but / does).
# didn't work in Mosaic 3.0 (but / does).
/ didn't work in Lynx 2.8 (but # does!)

Natko

7:33 pm on Apr 1, 2004 (gmt 0)

10+ Year Member



usefull reply R!, thanks.
I see new G-index, right now. I'm going to confront it and I'll be back with news.

ergophobe

4:38 pm on Apr 2, 2004 (gmt 0)

WebmasterWorld Senior Member 10+ Year Member Top Contributors Of The Month



Natko,

Here's my rationale

1. The savings are small (168 bytes on the page I tested). I live in a rural area where you can't get DSL so i don't assume that everyone does. We split 256KB ISDN connection between 2-4 computers. That's why I focus on the big stuff - optimizing images, CSS files, keeping images to a minimum and so on.

2. It is a standard, which means that complying user agents can be expected to observe the standard, but all bets are off if you choose to ignore the standard.

3. I can only test those user agents that are currently on the market or have been on the market in the past. I can't even test all of those. More importantly, though, I can't test the version that will come out tomorrow or three years from now. A strict standards page will not break just because someday a major user agent decides to enforce strict standards.

4. Sounds unlikely that a major user agnet will enforce strict standards? That's probably true as long as you serve up your pages as text/html mime type. However, I can't be sure that's how my pages will be served. I write most pages in XHTML 1.0 strict. Someone might decide in the future to serve up that page as application/xhtml+xml. In that case, guess what happens to all of my non-validating pages? They crash catastrophically and will display nothing except error messages. It's not that they will render poorly; they will not render at all.

5. In my non-web work, I work on editions that are printed exclusively with archival grade materials. They are meant to last at least 500 years - paper, ink, cover, binding should all be beautiful in 500 years if stored properly. The materials I work with are 450 years old.

Of course, I don't think on that time scale for my web pages, but I do tend to think of the long term and want to make sure that my web pages are *forward* compatible to the greates possible extent. IMO the best way to ensure that is to validate to a strict SGML DTD, since that is a relatively old and, therefore, well thought-out standard.

For me, saving 168 bytes just doesn't compare in value to that.

Tom

VectorJ

4:45 pm on Apr 2, 2004 (gmt 0)

10+ Year Member



It seems like the fraction of a second gained in load time by omiting 1 or 2 hundred bytes would probably be eaten up by the extra work the browser's parser would need to do. I assume that the first case that a parser checks for is properly formatted html, then if it can't find that it reparses the page and makes guesses.

That being said, I'm working on an SE crawler right now and the libraries I'm using (Perl LWP, HTML::Parser) have had no problem parsing questionably formatted html.

Natko

5:03 pm on Apr 2, 2004 (gmt 0)

10+ Year Member



Well, I'm begining to think that quoting #, hrefs and / its OK. But only these, nothing more and nothing less. Just these three. Maybe its ok to respect some recoms.
Anyway, I haven't seen any changes in position listings of my pages by new index on Google. I've remained top10 for those I was before. Even some new pages I added, linked bald, are indexed.
Maybe I'm not the only one not respecting all the recoms. :-)))
Well, we could say that even if quoted #, hrefs and/ are not necesary at all, these days, we better have to use them, becouse they are still part of w3c recoms, and this means that we don't know why to use tham.
A messy conclusion, but I'm not of english mother tongue. ;-)

ergophobe

5:57 pm on Apr 2, 2004 (gmt 0)

WebmasterWorld Senior Member 10+ Year Member Top Contributors Of The Month




Maybe I'm not the only one not respecting all the recoms. :-)))

As I said at the outset, SEs must *not* expect well-formed documents. It would be stupid for them to do so since their mission is not to validate documents, but to help people find content. The vast majority of documents out there are *not* well-formed HTML, so SEs mostly don't care.

I don't, however, jump from there to the conclusion that there is therefore no reason to create well-formed documents.

grahamstewart

6:37 pm on Apr 2, 2004 (gmt 0)

WebmasterWorld Senior Member 10+ Year Member



As I said at the outset, SEs must *not* expect well-formed documents.

Absolutely - a classic example of the Internet Engineering Task Force (IETF) fundamental rule:

"Be strict in what you send and generous in what you accept."

Wise words for anyone concerned with writing anything for the internet (including web pages!).

Natko

7:23 pm on Apr 2, 2004 (gmt 0)

10+ Year Member



I wanted to say that Google did't penalize my bald pages. No ponder in its algo for this!
This 42 message thread spans 2 pages: 42