Thread: PostgreSQL web site
Just a suggestion, why don't the web admins turn mod_gzip on? There are no downsides, and it would likely make a fairly large impact on bandwidth bills. Think of all the people browsing the PG documentation (myself included). Not only would bandwidth bills go down, responsiveness on the docs would be higher as well. Even if there is no bandwidth saving (if your b/w is free for instance), I'm of the opinion that mod_gzip and mod_deflate should be used unless there is a compelling reason not to. Sorry if this has been discussed before, but I thought it was worth mentioning on the off chance it hadn't. - Naz.
Naz Gassiep wrote:
> Just a suggestion, why don't the web admins turn mod_gzip on? There are
> no downsides, and it would likely make a fairly large impact on
> bandwidth bills. Think of all the people browsing the PG documentation
> (myself included). Not only would bandwidth bills go down,
> responsiveness on the docs would be higher as well.
>
> Even if there is no bandwidth saving (if your b/w is free for instance),
> I'm of the opinion that mod_gzip and mod_deflate should be used unless
> there is a compelling reason not to.
>
> Sorry if this has been discussed before, but I thought it was worth
> mentioning on the off chance it hadn't.
Actually that is a very good question... why are we not using mod_deflate?
Joshua D. Drake
>
> - Naz.
>
> ---------------------------(end of broadcast)---------------------------
> TIP 9: In versions below 8.0, the planner will ignore your desire to
> choose an index scan if your joining column's datatypes do not
> match
>
--
=== The PostgreSQL Company: Command Prompt, Inc. ===
Sales/Support: +1.503.667.4564 || 24x7/Emergency: +1.800.492.2240
Providing the most comprehensive PostgreSQL solutions since 1997
http://www.commandprompt.com/
> > Just a suggestion, why don't the web admins turn mod_gzip on? There > > are no downsides, and it would likely make a fairly large impact on > > bandwidth bills. Think of all the people browsing the PG > documentation > > (myself included). Not only would bandwidth bills go down, > > responsiveness on the docs would be higher as well. > > > > Even if there is no bandwidth saving (if your b/w is free for > > instance), I'm of the opinion that mod_gzip and mod_deflate > should be > > used unless there is a compelling reason not to. > > > > Sorry if this has been discussed before, but I thought it was worth > > mentioning on the off chance it hadn't. > > Actually that is a very good question... why are we not using > mod_deflate? Good question. It's definitly worth investigating. It does add load to the server, but not all that much. And if we have the space (need to check that on some of the boxen) we could set up pre-compressed static images and that argument would go away.. //Magnus
"Magnus Hagander" <mha@sollentuna.net> writes:
>> Actually that is a very good question... why are we not using
>> mod_deflate?
> Good question. It's definitly worth investigating.
What are we talking about here --- some hack to make users' web browsers
decompress pages on-the-fly? How much does that slow down the browsing
experience, if you've got an old slow PC? (I can believe that if you've
got a fast PC and a slow internet connection, it could make things
faster overall ... but the breakeven point is not obvious.) What are
the odds that people using older browsers will be locked out entirely?
regards, tom lane
Also,
I'd like to see PDA version of web site and archives. Current design
is very inconvenient.
Oleg
On Mon, 28 Aug 2006, Magnus Hagander wrote:
>>> Just a suggestion, why don't the web admins turn mod_gzip on? There
>>> are no downsides, and it would likely make a fairly large impact on
>>> bandwidth bills. Think of all the people browsing the PG
>> documentation
>>> (myself included). Not only would bandwidth bills go down,
>>> responsiveness on the docs would be higher as well.
>>>
>>> Even if there is no bandwidth saving (if your b/w is free for
>>> instance), I'm of the opinion that mod_gzip and mod_deflate
>> should be
>>> used unless there is a compelling reason not to.
>>>
>>> Sorry if this has been discussed before, but I thought it was worth
>>> mentioning on the off chance it hadn't.
>>
>> Actually that is a very good question... why are we not using
>> mod_deflate?
>
> Good question. It's definitly worth investigating. It does add load to
> the server, but not all that much. And if we have the space (need to
> check that on some of the boxen) we could set up pre-compressed static
> images and that argument would go away..
>
> //Magnus
>
> ---------------------------(end of broadcast)---------------------------
> TIP 1: if posting/reading through Usenet, please send an appropriate
> subscribe-nomail command to majordomo@postgresql.org so that your
> message can get through to the mailing list cleanly
>
Regards,
Oleg
_____________________________________________________________
Oleg Bartunov, Research Scientist, Head of AstroNet (www.astronet.ru),
Sternberg Astronomical Institute, Moscow University, Russia
Internet: oleg@sai.msu.su, http://www.sai.msu.su/~megera/
phone: +007(495)939-16-83, +007(495)939-23-83
<pre wrap="">
</pre><blockquote cite="mid15501.1156790439@sss.pgh.pa.us" type="cite"><pre wrap="">
What are we talking about here --- some hack to make users' web browsers
decompress pages on-the-fly? How much does that slow down the browsing
experience, if you've got an old slow PC? (I can believe that if you've
got a fast PC and a slow internet connection, it could make things
faster overall ... but the breakeven point is not obvious.) What are
the odds that people using older browsers will be locked out entirely?
regards, tom lane</pre></blockquote> I have been using mod_deflate for years on my web server, and have yet to
notice_any_ difference on even very old hardware. Furthermore, the additional load to the server is quite small.<br
/><br/> If you are concerned about widespread impact on users, consider that slashdot and google are both
mod_deflated.<br/><br /> - Naz.<br />
<br /><blockquote cite="mid15501.1156790439@sss.pgh.pa.us" type="cite"><blockquote type="cite"><pre wrap="">Good
question.It's definitly worth investigating. </pre></blockquote><pre wrap="">
What are we talking about here --- some hack to make users' web browsers
decompress pages on-the-fly? How much does that slow down the browsing
experience, if you've got an old slow PC? (I can believe that if you've
got a fast PC and a slow internet connection, it could make things
faster overall ... but the breakeven point is not obvious.) What are
the odds that people using older browsers will be locked out entirely?
regards, tom lane
</pre></blockquote> Oh, and there is no chance that older browsers are locked out, mod_deflate and mod_gzip (gzip is
theapache 1.3 module, deflate is for apache 2) both only compress if they find that the browser supports it. That being
said,gzipped content was supported by browsers since the netscape 3 days.<br /> Regards,<br /> - Naz<br />
> >> Actually that is a very good question... why are we not using > >> mod_deflate? > > > Good question. It's definitly worth investigating. > > What are we talking about here --- some hack to make users' > web browsers decompress pages on-the-fly? Yes. > How much does that > slow down the browsing experience, if you've got an old slow > PC? (I can believe that if you've got a fast PC and a slow > internet connection, it could make things faster overall ... > but the breakeven point is not obvious.) In most cases it's faster or unnoticably slower. At the speed your machine would have to be to notice the problem, I don't think you can load a modern-enough browser to actually look at the site. > What are the odds > that people using older browsers will be locked out entirely? They should be zero, if you use mod_deflate. The browser sends an Accept-encoding: deflate (or similar, I'm not 100% on the name of the header, and too lazy to look it up). If this is not present in the request, an uncompressed version of the page is sent back to the browser. //Magnus
Magnus Hagander wrote: > Good question. It's definitly worth investigating. It does add load to > the server, but not all that much. And if we have the space (need to > check that on some of the boxen) we could set up pre-compressed static > images and that argument would go away.. > > //Magnus > > ---------------------------(end of broadcast)--------------------------- > > I generally only compress HTML and other text formats ignoring images, PDFs and other binary data. I would think this would have a huge impact on the PG site, which would deliver a lot of text serving documentation Regards, - Naz
>
> Good question. It's definitly worth investigating. It does add load to
> the server, but not all that much. And if we have the space (need to
> check that on some of the boxen) we could set up pre-compressed static
> images and that argument would go away..
we can configure deflate so that it only compresses the html not the images.
Joshua D. Drake
>
> //Magnus
>
--
=== The PostgreSQL Company: Command Prompt, Inc. ===
Sales/Support: +1.503.667.4564 || 24x7/Emergency: +1.800.492.2240
Providing the most comprehensive PostgreSQL solutions since 1997
http://www.commandprompt.com/
Tom Lane wrote:
> "Magnus Hagander" <mha@sollentuna.net> writes:
>>> Actually that is a very good question... why are we not using
>>> mod_deflate?
>
>> Good question. It's definitly worth investigating.
>
> What are we talking about here --- some hack to make users' web browsers
> decompress pages on-the-fly?
No it is not a hack. With HTTP1.1 browers have the ability to compress
and decompress on the fly. This is done in the negotiation between
> How much does that slow down the browsing
> experience, if you've got an old slow PC?
Well good question. How slow are we talking? If we are talking a AMD 300
people are likely not using broadband and it wouldn't matter anyway
because either way they are use to slow.
> (I can believe that if you've
> got a fast PC and a slow internet connection, it could make things
> faster overall ... but the breakeven point is not obvious.) What are
> the odds that people using older browsers will be locked out entirely?
None. Because it will fall back to 1.0 if your browser doesn't support
it. The browser will negotiate with the web server to determine the best
route to retrieve the data.
Sincerely,
Joshua D. Drake
>
> regards, tom lane
>
--
=== The PostgreSQL Company: Command Prompt, Inc. ===
Sales/Support: +1.503.667.4564 || 24x7/Emergency: +1.800.492.2240
Providing the most comprehensive PostgreSQL solutions since 1997
http://www.commandprompt.com/
Oleg Bartunov wrote: > Also, > > I'd like to see PDA version of web site and archives. Current design is > very inconvenient. Egad... what a pain. Please feel free to work one up for us, but a PDA version is pretty much a pain in the arse for little benefit. Sincerely, Josuha D. Drake > > Oleg > On Mon, 28 Aug 2006, Magnus Hagander wrote: > >>>> Just a suggestion, why don't the web admins turn mod_gzip on? There >>>> are no downsides, and it would likely make a fairly large impact on >>>> bandwidth bills. Think of all the people browsing the PG >>> documentation >>>> (myself included). Not only would bandwidth bills go down, >>>> responsiveness on the docs would be higher as well. >>>> >>>> Even if there is no bandwidth saving (if your b/w is free for >>>> instance), I'm of the opinion that mod_gzip and mod_deflate >>> should be >>>> used unless there is a compelling reason not to. >>>> >>>> Sorry if this has been discussed before, but I thought it was worth >>>> mentioning on the off chance it hadn't. >>> >>> Actually that is a very good question... why are we not using >>> mod_deflate? >> >> Good question. It's definitly worth investigating. It does add load to >> the server, but not all that much. And if we have the space (need to >> check that on some of the boxen) we could set up pre-compressed static >> images and that argument would go away.. >> >> //Magnus >> >> ---------------------------(end of broadcast)--------------------------- >> TIP 1: if posting/reading through Usenet, please send an appropriate >> subscribe-nomail command to majordomo@postgresql.org so that your >> message can get through to the mailing list cleanly >> > > Regards, > Oleg > _____________________________________________________________ > Oleg Bartunov, Research Scientist, Head of AstroNet (www.astronet.ru), > Sternberg Astronomical Institute, Moscow University, Russia > Internet: oleg@sai.msu.su, http://www.sai.msu.su/~megera/ > phone: +007(495)939-16-83, +007(495)939-23-83 > > ---------------------------(end of broadcast)--------------------------- > TIP 9: In versions below 8.0, the planner will ignore your desire to > choose an index scan if your joining column's datatypes do not > match > -- === The PostgreSQL Company: Command Prompt, Inc. === Sales/Support: +1.503.667.4564 || 24x7/Emergency: +1.800.492.2240 Providing the most comprehensive PostgreSQL solutions since 1997 http://www.commandprompt.com/
Joshua D. Drake Wrote > we can configure deflate so that it only compresses the html > not the images. I think he meant, pre compressed html pages. Eg,: index.html index.html.gz etc... ... John
<br /> John Hansen wrote: <blockquote cite="midC88CF37B2468134A9A28F4B2BFDEBAB901A663@bdc.geeknet.com.au" type="cite"><prewrap="">Joshua D. Drake Wrote </pre><blockquote type="cite"><pre wrap="">we can configure deflate so that it only compresses the html not the images. </pre></blockquote><pre wrap=""> I think he meant, pre compressed html pages. Eg,: index.html index.html.gz etc... </pre></blockquote> Oh, thats a bad idea, I think, as it would hugely increase the maintenance work on the site, for a marginalgain in server performance. The load increase really is minimal. If our servers are so close to breaking point thatthis is an issue then we'll need new ones very soon anyway.<br /> Regards,<br /> - Naz<br />
Tom Lane Wrote: >"Magnus Hagander" <mha@sollentuna.net> writes: >>> Actually that is a very good question... why are we not using >>> mod_deflate? > >> Good question. It's definitly worth investigating. > > What are we talking about here --- some hack to make users' web browsers decompress pages on-the-fly? How much > doesthat slow down the browsing experience, if you've got an old slow PC? (I can believe that if you've got a > fast PCand a slow internet connection, it could make things faster overall ... but the breakeven point is not > obvious.) What are the odds that people using older browsers will be locked out entirely? Documents are only compressed if the browser requests one, so nobody would get locked out. And for the odd users with a slow PC, they would probably already have cofigured their browser so it doesn't request compresseddocuments. Kind Regards, John
>> index.html
>> index.html.gz
>>
>> etc...
>>
>>
> Oh, thats a bad idea, I think, as it would hugely increase the maintenance work
> on the site, for a marginal gain in server performance. The load increase really
> is minimal. If our servers are so close to breaking point that this is an issue
> then we'll need new ones very soon anyway.
Yeah... no. Mod_deflate will allow auto-negotiation so that the
information can be dealt with either way. No reason to increase the
workload.
> Regards,
> - Naz
--
=== The PostgreSQL Company: Command Prompt, Inc. ===
Sales/Support: +1.503.667.4564 || 24x7/Emergency: +1.800.492.2240
Providing the most comprehensive PostgreSQL solutions since 1997
http://www.commandprompt.com/
"Magnus Hagander" <mha@sollentuna.net> writes:
>> What are the odds
>> that people using older browsers will be locked out entirely?
> They should be zero, if you use mod_deflate.
> The browser sends an Accept-encoding: deflate (or similar, I'm not 100%
> on the name of the header, and too lazy to look it up). If this is not
> present in the request, an uncompressed version of the page is sent back
> to the browser.
Got it. Thanks for the explanation.
regards, tom lane
> >> index.html > >> index.html.gz > >> > >> etc... > >> > >> > > Oh, thats a bad idea, I think, as it would hugely increase the > > maintenance work on the site, for a marginal gain in server > > performance. The load increase really is minimal. If our > servers are > > so close to breaking point that this is an issue then we'll > need new ones very soon anyway. > > Yeah... no. Mod_deflate will allow auto-negotiation so that > the information can be dealt with either way. No reason to > increase the workload. We're nowhere near the load that our static servers can deliver. Even during release time, when it increases sometimes hundredfold. That doesn't mean we shouldn't optimize when we can do so at trivial cost, though. //Magnus
> We're nowhere near the load that our static servers can deliver. Even
> during release time, when it increases sometimes hundredfold.
> That doesn't mean we shouldn't optimize when we can do so at trivial
> cost, though.
Well more specifically, this will be a great boon for lowering bandwidth
requirements on both sides of the content delivery.
Joshua D. Drake
>
> //Magnus
>
--
=== The PostgreSQL Company: Command Prompt, Inc. ===
Sales/Support: +1.503.667.4564 || 24x7/Emergency: +1.800.492.2240
Providing the most comprehensive PostgreSQL solutions since 1997
http://www.commandprompt.com/
-----Original Message----- From: "Joshua D. Drake" <jd@commandprompt.com> To: "Oleg Bartunov" <oleg@sai.msu.su> Cc: "Magnus Hagander" <mha@sollentuna.net>; "Naz Gassiep" <naz@mira.net>; "pgsql-www@postgresql.org" <pgsql-www@postgresql.org> Sent: 28/08/06 20:19 Subject: Re: [pgsql-www] PostgreSQL web site > a pain in the arse for little benefit. Wow, never seen a left-pondian spell it that way - are you secretly a Brit or Aussie? /D
"Joshua D. Drake" <jd@commandprompt.com> writes:
> we can configure deflate so that it only compresses the html not the images.
One would certainly hope so, since the images are compressed already.
regards, tom lane
On Mon, 28 Aug 2006, Joshua D. Drake wrote: > Oleg Bartunov wrote: >> Also, >> >> I'd like to see PDA version of web site and archives. Current design is >> very inconvenient. > > Egad... what a pain. Please feel free to work one up for us, but a PDA > version is pretty much a pain in the arse for little benefit. As a first step, changing CSS optimized for PDA would be ok. Archives is an our knowledgebase and sometimes I need to search/read them using my PDA (it's 640x480). But I don't insist :) > > Sincerely, > > Josuha D. Drake > > >> >> Oleg >> On Mon, 28 Aug 2006, Magnus Hagander wrote: >> >>>>> Just a suggestion, why don't the web admins turn mod_gzip on? There >>>>> are no downsides, and it would likely make a fairly large impact on >>>>> bandwidth bills. Think of all the people browsing the PG >>>> documentation >>>>> (myself included). Not only would bandwidth bills go down, >>>>> responsiveness on the docs would be higher as well. >>>>> >>>>> Even if there is no bandwidth saving (if your b/w is free for >>>>> instance), I'm of the opinion that mod_gzip and mod_deflate >>>> should be >>>>> used unless there is a compelling reason not to. >>>>> >>>>> Sorry if this has been discussed before, but I thought it was worth >>>>> mentioning on the off chance it hadn't. >>>> >>>> Actually that is a very good question... why are we not using >>>> mod_deflate? >>> >>> Good question. It's definitly worth investigating. It does add load to >>> the server, but not all that much. And if we have the space (need to >>> check that on some of the boxen) we could set up pre-compressed static >>> images and that argument would go away.. >>> >>> //Magnus >>> >>> ---------------------------(end of broadcast)--------------------------- >>> TIP 1: if posting/reading through Usenet, please send an appropriate >>> subscribe-nomail command to majordomo@postgresql.org so that your >>> message can get through to the mailing list cleanly >>> >> >> Regards, >> Oleg >> _____________________________________________________________ >> Oleg Bartunov, Research Scientist, Head of AstroNet (www.astronet.ru), >> Sternberg Astronomical Institute, Moscow University, Russia >> Internet: oleg@sai.msu.su, http://www.sai.msu.su/~megera/ >> phone: +007(495)939-16-83, +007(495)939-23-83 >> >> ---------------------------(end of broadcast)--------------------------- >> TIP 9: In versions below 8.0, the planner will ignore your desire to >> choose an index scan if your joining column's datatypes do not >> match >> > > > Regards, Oleg _____________________________________________________________ Oleg Bartunov, Research Scientist, Head of AstroNet (www.astronet.ru), Sternberg Astronomical Institute, Moscow University, Russia Internet: oleg@sai.msu.su, http://www.sai.msu.su/~megera/ phone: +007(495)939-16-83, +007(495)939-23-83
Dave Page wrote:
>
> -----Original Message-----
> From: "Joshua D. Drake" <jd@commandprompt.com>
> To: "Oleg Bartunov" <oleg@sai.msu.su>
> Cc: "Magnus Hagander" <mha@sollentuna.net>; "Naz Gassiep" <naz@mira.net>; "pgsql-www@postgresql.org"
<pgsql-www@postgresql.org>
> Sent: 28/08/06 20:19
> Subject: Re: [pgsql-www] PostgreSQL web site
>
>> a pain in the arse for little benefit.
>
> Wow, never seen a left-pondian spell it that way - are you secretly a Brit or Aussie?
Heh no I am pure American, which means I am everything and nothing. I
use that spelling when I want to swear without swearing. Similar to the
use of "bloody".
Sincerely,
Joshua D. Drake
>
> /D
> ---------------------------(end of broadcast)---------------------------
> TIP 2: Don't 'kill -9' the postmaster
>
--
=== The PostgreSQL Company: Command Prompt, Inc. ===
Sales/Support: +1.503.667.4564 || 24x7/Emergency: +1.800.492.2240
Providing the most comprehensive PostgreSQL solutions since 1997
http://www.commandprompt.com/
Hello, On Tue, 2006-08-29 at 00:28 +0400, Oleg Bartunov wrote: > > Egad... what a pain. Please feel free to work one up for us, but a > > PDAversion is pretty much a pain in the arse for little benefit. > > As a first step, changing CSS optimized for PDA would be ok. Archives > is an our knowledgebase and sometimes I need to search/read them using > my PDA (it's 640x480). I want a lynx and links version of the PostgreSQL website, please ;) Cheers, -- The PostgreSQL Company - Command Prompt, Inc. 1.503.667.4564 PostgreSQL Replication, Consulting, Custom Development, 24x7 support Managed Services, Shared and Dedicated Hosting Co-Authors: plPHP, plPerlNG - http://www.commandprompt.com/
Attachment
> -----Original Message----- > From: pgsql-www-owner@postgresql.org > [mailto:pgsql-www-owner@postgresql.org] On Behalf Of Devrim GUNDUZ > Sent: 29 August 2006 15:22 > To: Oleg Bartunov > Cc: Joshua D. Drake; Magnus Hagander; Naz Gassiep; > pgsql-www@postgresql.org > Subject: Re: [pgsql-www] PostgreSQL web site > > Hello, > > On Tue, 2006-08-29 at 00:28 +0400, Oleg Bartunov wrote: > > > Egad... what a pain. Please feel free to work one up for > us, but a > > > PDAversion is pretty much a pain in the arse for little benefit. > > > > As a first step, changing CSS optimized for PDA would be > ok. Archives > > is an our knowledgebase and sometimes I need to search/read > them using > > my PDA (it's 640x480). > > I want a lynx and links version of the PostgreSQL website, please ;) <joke tagtype="just to make sure it's obvious">And I'd like a video of the conference highlights</joke> :-) /D
Hello, On Tue, 2006-08-29 at 15:44 +0100, Dave Page wrote: > > I want a lynx and links version of the PostgreSQL website, please ;) > > <joke tagtype="just to make sure it's obvious">And I'd like a video of > the conference highlights</joke> /me hides :-D -- The PostgreSQL Company - Command Prompt, Inc. 1.503.667.4564 PostgreSQL Replication, Consulting, Custom Development, 24x7 support Managed Services, Shared and Dedicated Hosting Co-Authors: plPHP, plPerlNG - http://www.commandprompt.com/
Attachment
> Heh no I am pure American, which means I am everything and nothing. I > use that spelling when I want to swear without swearing. Similar to > the use of "bloody". > It's actually supposed to be the other way around hehe. "Ass" is the old English word for donkey, while "arse" is the reference to one's buttocks. So describing someone as an "ass" could really be argued to be quite proper even in heated conversation, as it could be a reference to their stubborn nature or unwillingness to compromise, while calling another an "arse" is an unambiguous derogatory linking between them and a posterior.