W3C home > Mailing lists > Public > site-comments@w3.org > August 2009

RE: compression of HTML would save a lot of money

From: Rotan Hanrahan <rotan.hanrahan@mobileaware.com>
Date: Tue, 4 Aug 2009 13:39:45 +0100
Message-ID: <D5306DC72D165F488F56A9E43F2045D301FF4660@FTO.mobileaware.com>
To: "FinanzNachrichten.de, Markus Meister" <markus.meister@finanznachrichten.de>, <site-comments@w3.org>, <w3c-ac-forum@w3.org>
Hi Markus, 

It's not just IE, there are plenty others with issues, but that's not the real problem.

Think global.

In one fell swoop, the W3C could block access to its vast Web resources to many thousands (if not millions) of users in under-developed and developing countries, many of whom would lack the IT resources to understand and compensate for the change. So what if they are a mere 1%, or 0.1% or whatever? The W3C is in a unique position; its mission to lead the Web to its full potential needs to be as inclusive as possible, especially to those who are less fortunate.

Don't get me wrong, I am not objecting to the use of compression per se, just its blanket application.

I propose an alternative strategy:

1. Do not compress the home page (and include a link to a non-compressed page that explains the gzip decision and explains a means of reporting problems that have resulted from the introduction of gzip).

2. If the HTTP request does not indicate gzip support then obviously don't gzip the response.

3. If the UA header matches a pattern of known "fully gzip supporting" browsers, then gzip the response.

This is an "opt in" strategy. It is perhaps the safest approach. At the very start, no browsers will receive gzip responses, which is what we have today. As browsers are identified that fully support gzip, and which can be detected from evidence in the HTTP request, then more and more of the Web users will receive gzip responses. Obviously we could address a sizeable portion of the community with just a handful of the latest browsers, and that may be enough to get the traffic/cost benefits that gzip could offer.

Originally I thought of possibly using an "opt out" approach, where "problem" browsers would be singled out for uncompressed responses, but the initial negative effect on the Web community might cause unnecessary problems and tarnish the good reputation of the consortium.

---Rotan.

-----Original Message-----
From: FinanzNachrichten.de, Markus Meister [mailto:markus.meister@finanznachrichten.de] 
Sent: 04 August 2009 12:24
To: Rotan Hanrahan; site-comments@w3.org; w3c-ac-forum@w3.org
Subject: AW: compression of HTML would save a lot of money

Good hint, Rotan.
I just had a look at the traffic stats of our media.
Last month we had 0,06 % of the traffic with IE5.5 and a total of 0,05 %
with IE5/IE5.01/IE5.14 and older versions of IE.
That's in total around 0,11 %.

Sounds tough, but I would prefer to 'motivate' this 1 out of 1000 users to
update his browser ...

Markus



-----Ursprüngliche Nachricht-----
Von: w3c-ac-forum-request@w3.org [mailto:w3c-ac-forum-request@w3.org] Im
Auftrag von Rotan Hanrahan
Gesendet: Dienstag, 4. August 2009 12:48
An: FinanzNachrichten.de, Markus Meister; site-comments@w3.org;
w3c-ac-forum@w3.org
Betreff: RE: compression of HTML would save a lot of money

Not all browsers support gzip, mainly older ones. Some browsers that claim
to support gzip have bugs (e.g. IE5.5 without the hotfix) that cause page
loading to fail or worse. The W3C server would therefore have to "sniff" the
requests to avoid causing problems for a subset of the global community. I'm
not sure if the hotfix can be detected in a HTTP request so we'd have to
take the safest line and not deliver gzip to any of these browsers, fixed or
otherwise. Even some of the newer browsers have problems with gzip.

Compression would also increase the processing overhead a little, though
caching could help here.

So, while I think the proposal has merit, we need to be cautious that we
don't detach a portion of the world who are struggling with old and/or buggy
browser implementations. W3C is global, after all.

---Rotan



-----Original Message-----
From: w3c-ac-forum-request@w3.org [mailto:w3c-ac-forum-request@w3.org] On
Behalf Of FinanzNachrichten.de, Markus Meister
Sent: 04 August 2009 11:29
To: site-comments@w3.org; w3c-ac-forum@w3.org
Subject: compression of HTML would save a lot of money

Dear Ian,

I have just noticed that we don't use HTML compression on our W3C website.
If we would use e.g. GZIP, we could save a lot of traffic and money.

3 examples:

http://www.w3.org/       46 kB
http://lists.w3.org/    227 kB
http://validator.w3.org/ 27 kB

I have just put these three html files on our test server and the result
confirmed that if we would deliver compressed files, the size would somewhat
smaller and we could reduce our costs by saving cash:

* homepage compressed:          14 kB (-70 % traffic)
* lists homepage compressed:    23 kB (-90 % traffic)
* validator homepage compressed: 5 kB (-81 % traffic)


Looking at our financial situation, it would be phantastic if you could
reduce our traffic expenses by maybe -50 % (graphics would not be concerned
by a compression).

If there are no 'political' reasons why it's not possible for us to use
compression, please check this idea with the server responsible.
When the traffic numbers are still around the same as last year (when I
proposed advertising to increase our revenues) I would guess that we could
save as much traffic costs that we economize a 5 digit Euro/USD amount per
year with compression.

Best regards,

FINANZNACHRICHTEN.DE
Markus Meister

-------------------------------------------
http://www.finanznachrichten.de
Alle News zu Aktien, Börse und Finanzen!!
-------------------------------------------
DER SPEKULANT - Der Börsenbrief für clevere Anleger 
http://www.derspekulant.ch
-------------------------------------------
Received on Tuesday, 4 August 2009 12:40:40 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Wednesday, 24 October 2012 16:21:31 GMT