alexander.kirk.at— Not the usual guide suggesting to strip all white space. This guide speaks about generation time of sites and how to optimize it.
Feb 2, 2006View in Crawl 4
Here's my suggestion: Don't run a database-driven website when all you have is static content. What's wrong with plain jane HTML for simple pages? Not every web page has to have all the latest bells & whistles & RSS feeds and all that.
@nicholasjon,> To declare either HTML or XHTML "crap" outright is pretty useless. Just choose the right tool for you and the job you're doing.I agree.> For me, that tool is XHTML -- because I like following the strict DOCTYPEHTML 4.01 has a Strict doctype.> I like being able to identify problems with a validatorYou can validate HTML 4.01 just as easily as you can validate XHTML.> and I like knowing the various browsers are using "standards mode" to render my pages.You can use "standards mode" with HTML 4.01 just as easily as with XHTML. In fact, the XML prologue many people use with XHTML kicks some browsers out of "standards mode".@alexanderkirk,> i prefer xhtml because it's an up-to-date standard.Larger version numbers don't automatically translate into better choices. For instance, do you use XHTML 1.1, even though it's completely incompatible with Internet Explorer? Or do you use the less up-to-date XHTML 1.0?> you want xml compatibility with your html pages.Unless you are serving it as application/xhtml+xml or similar, you aren't getting XML compatibility.> but if you choose to take html4, ok, but i think arguing it with the "less size" is useless.I don't think the lesser size is very important either, I just think it's stupid for an article that is supposed to be telling you how to reduce page size explicitly recommends that you switch to a worse choice.
Dugg. This is the best web site optimization article I've read yet, in that it has only a handful of glaring problems.He doesn't talk about latency enough -- persistent HTTP connections make a huge difference. DNS plays a role.With a reverse proxy, component cache is redundant.No mention of database connection pooling. Maybe because PHP doesn't let you do it.That Last-Modified header won't make your page usable by shared caches if you've set a cookie. Using cookies judiciously should be covered.Still, lots of important tips were in there that more people should know about.
> With a reverse proxy, component cache is redundant.Not so. It works best when some parts of the page are expensive to create and are updated infrequently and some parts of the page are cheap to create and are updated frequently. A reverse proxy won't help in this situation, because you need to keep up with the frequent updates of the cheap components, and the reverse proxy can't cache different parts of the page independently.
Closed AccountFeb 2, 2006
Here's my suggestion: Don't run a database-driven website when all you have is static content. What's wrong with plain jane HTML for simple pages? Not every web page has to have all the latest bells & whistles & RSS feeds and all that.
bogthaFeb 2, 2006
What is so "crappy" about HTML 4? It's a better choice than XHTML for most people.
bogthaFeb 3, 2006
@nicholasjon,> To declare either HTML or XHTML "crap" outright is pretty useless. Just choose the right tool for you and the job you're doing.I agree.> For me, that tool is XHTML -- because I like following the strict DOCTYPEHTML 4.01 has a Strict doctype.> I like being able to identify problems with a validatorYou can validate HTML 4.01 just as easily as you can validate XHTML.> and I like knowing the various browsers are using "standards mode" to render my pages.You can use "standards mode" with HTML 4.01 just as easily as with XHTML. In fact, the XML prologue many people use with XHTML kicks some browsers out of "standards mode".@alexanderkirk,> i prefer xhtml because it's an up-to-date standard.Larger version numbers don't automatically translate into better choices. For instance, do you use XHTML 1.1, even though it's completely incompatible with Internet Explorer? Or do you use the less up-to-date XHTML 1.0?> you want xml compatibility with your html pages.Unless you are serving it as application/xhtml+xml or similar, you aren't getting XML compatibility.> but if you choose to take html4, ok, but i think arguing it with the "less size" is useless.I don't think the lesser size is very important either, I just think it's stupid for an article that is supposed to be telling you how to reduce page size explicitly recommends that you switch to a worse choice.
brlewisFeb 3, 2006
Dugg. This is the best web site optimization article I've read yet, in that it has only a handful of glaring problems.He doesn't talk about latency enough -- persistent HTTP connections make a huge difference. DNS plays a role.With a reverse proxy, component cache is redundant.No mention of database connection pooling. Maybe because PHP doesn't let you do it.That Last-Modified header won't make your page usable by shared caches if you've set a cookie. Using cookies judiciously should be covered.Still, lots of important tips were in there that more people should know about.
bogthaFeb 3, 2006
> With a reverse proxy, component cache is redundant.Not so. It works best when some parts of the page are expensive to create and are updated infrequently and some parts of the page are cheap to create and are updated frequently. A reverse proxy won't help in this situation, because you need to keep up with the frequent updates of the cheap components, and the reverse proxy can't cache different parts of the page independently.