Search engine optimization hints

Web-lay-out suggestions, functionality, reporting dead links etc
halfgaar
Posts: 150
Joined: Sat 23 Feb 2008 19:48
Location: The Netherlands

Search engine optimization hints

Postby halfgaar » Sun 30 Mar 2008 20:53

Hi,

I've got a couple of remarks about the search engine friendliness of waisays.com, and related sites. Because to be frank, it's pretty bad as it is now. The fact that the Google query "acne raw food" doesn't even yield anything Wai related on the first 4 pages, proves it.

So, I'd like to give a view tips for better SEO (and some user friendliness as well):

1) There are several domain names in use now. I found freeacnebook.com, waisays.com, youngerthanyourage.com, waicure.com and waidiet.com with some short searching. This is bad for multiple reasons. One is that your incoming links all point to different pages, so the total score you can attain is spread over multiple sites, reducing the score per site. Another is that there is duplicate content on some of them. Duplicate content on different websites is punished by search engines (i.e. you loose ranking).

To solve this, let all domains do an HTTP 301 redirect to one single address. This also means that waisays.com should redirect to www.waisays.com, because they are two different sites from a technical point of view, causing 100% duplicate content.

2) Some pages are only found by deeplinking (meaning, links somewhere in an article, as opposed to in some main menu. This is mostly user unfriendly, as opposed to search engine unfriendly, however, there is also a search engine issue here, and that is that the menu's are javascript based. Search engines don't use javascript, so they don't see the links. If you really want to maintain the javascript links (which I recommend against), make a sitemap which contains all the pages of the site, and put that in a robots.txt.

3) Don't use frames. Some robots may understand them, but Google for example, recommends against them.

4) Make your link texts meaningful. A link text like "more" or "here" (from "click here") doesn't describe the page you're linking to. Better would be make the descriptive text the link, as I do in this forum message.

Unfortunately, a lot of this requires major rewriting and rethinking, but it should be worth it. Feel free to ignore it, but I though I'd mention it :)

For further reading, you may wanna read the Google Webmaster Guidelines.
User avatar
RRM
Administrator
Posts: 8153
Joined: Sat 16 Jul 2005 00:01
Contact:

Postby RRM » Wed 09 Apr 2008 18:41

Thank you halfgaar.
Unfortunately i dont have time at the moment to change anything. And neither does Oscar, im afraid.
Do you have some time on your hands?

BTW, do you do the diet somewhat (like 50%), and thats why your name is halfgaar? Or is it because you are proud of your independent thinking?
(for those who dont speak dutch, it means "half cooked" or "medium-rare", but it also means "weirdo")
halfgaar
Posts: 150
Joined: Sat 23 Feb 2008 19:48
Location: The Netherlands

Postby halfgaar » Wed 09 Apr 2008 19:02

Do you have some time on your hands?
To a certain extent, but continuing on what somebody else made, often involves a lot of work, and often incompatible design philosophies arise. For example, I'd engineer it with some XSLT and Ruby, on Linux, like I do for my own website, but that would make administrating for most people a pain :). Likewise, not working with it, would make it a pain for me :)

If you could give me a copy of all the contents (minus things that should not be seen perhaps), I could make a quick-and-dirty test-site. Can't say when I'll have it finished, though. I'll be sure to keep it out of search engines with a proper robots.txt.
BTW, do you do the diet somewhat (like 50%), and thats why your name is halfgaar? Or is it because you are proud of your independent thinking?
Well, I guess my nickname could be seen as a premonition. I chose it about 10 years ago, when I neither ate raw food, or did much independent thinking. But I have become an independent thinker, of which I'm proud indeed. One friend in particular calls my non-conformism and confidence in my controversial opinions, and willingness to stick by them if I believe in them, quite inspirational :)

And, recently, my nick has become more relevant because of my raw diet :). I am 100% raw at the moment, but plan to try some cooked foods once I'm acne free (enough). The pimples don't go away as fast as with most people, but it seems to be working. Also, I've got the flu now (a lot of people here do, so I hear), so I can imagine that influences it as well.

I started a thread on my progress (which has gotten no replies :(), so if you have something to add, please do :).
User avatar
RRM
Administrator
Posts: 8153
Joined: Sat 16 Jul 2005 00:01
Contact:

Postby RRM » Tue 29 Apr 2008 19:53

halfgaar wrote:If you could give me a copy of all the contents (minus things that should not be seen perhaps), I could make a quick-and-dirty test-site.
Sounds good.
What do you think Oscar?
halfgaar
Posts: 150
Joined: Sat 23 Feb 2008 19:48
Location: The Netherlands

Postby halfgaar » Tue 29 Apr 2008 21:44

As I said, I will be working with Linux, Ruby and XSLT to generate the website. Should I build something for it, and at some point not be able to do the webmastering, are there other people with a Linux machine which can take over?

But perhaps giving me the copy is a good start. It also allows me to see how it's done now. So what I need, is the entire working directory; not just what is uploaded (that is, if there is a difference).
User avatar
Oscar
Administrator
Posts: 4347
Joined: Mon 15 Aug 2005 00:01

Postby Oscar » Thu 01 May 2008 11:56

I'm a complete Linux ignorant, so taking over would be hard to do...

All help is welcome though, since, like RRM mentioned, we lack the time.

I'll make you a moderator of the New Site/Forum section, so you can read about the ideas and post your findings. RRM can give you access to the working directories. Thanks! :)
Havas
Posts: 51
Joined: Mon 14 Jul 2008 03:59

Postby Havas » Sun 10 Aug 2008 09:24

I don't mean this in a "I will only help you, if you help me" way, but if this diet is all it is made up to be, I would love to be able to help spread it.

So if I do get clear I would really appreciate if you guys let me give back. I have two development SuSE (Linux) servers. And a production server that I could (temporarily) host this site if need be.

Although, I would only use standard tools/libraries so that hosting is not an issue and can be taken anywhere (including the current server).

I've got extensive experience in PHP/MySQL/C++/C# and a working knowledge of windows and linux programming. And basic C (Without the libs). I'm currently working on the back end of stock website.


Halfgaar, I'm not sure why you'd want to use Ruby? It's not standard, its slow as molasses(PHP is 3 - 9 times the speed: http://shootout.alioth.debian.org/gp4/b ... &lang2=php). And using XSLT would be completely reckless in my opinion. I think you'd find trouble finding a comparable site that needs it. :)



Anyway, most importantly I have fair bit of time. (Well, perhaps I won't with clear skin ;D).

But honestly, I would really appreciate if I could do that if no one else does. I believe I could improve the site many times over. And ultimately help A LOT more people see.
halfgaar
Posts: 150
Joined: Sat 23 Feb 2008 19:48
Location: The Netherlands

Postby halfgaar » Sun 10 Aug 2008 11:27

What I was suggesting, is a static site, which is processed on the webmasters PC with Ruby and XSLT, and then uploaded as plain (X)HTML to the webserver. So the slowness is not really an issue. The XSLT would be necessary to add layout, headers, and footers to plain pages, to avoid redundancy.

But I think a better way would be to convert the site to a simple PHP content management system, like Wordpress. However, Wordpress does take more effort on the webmaster's part in maintaining, as upgrading webapps is a pain. And the webmaster would need to be experienced with PHP, MySQL, diff+patch (preferably) and perhaps some more.

As for the hosting. We don't need to change server, I guess. And if we do, I suggest we move over to Nearly Free Speech. Any tech lover loves it :)

But can we have an overview of which sites belong to "Wai"? And could we perhaps have access to the source files? Then we can begin planning some changes.
Havas
Posts: 51
Joined: Mon 14 Jul 2008 03:59

Postby Havas » Sun 10 Aug 2008 19:46

halfgaar wrote:What I was suggesting, is a static site, which is processed on the webmasters PC with Ruby and XSLT, and then uploaded as plain (X)HTML to the webserver. So the slowness is not really an issue. The XSLT would be necessary to add layout, headers, and footers to plain pages, to avoid redundancy.
While technically it works, it's not what it was designed for. XSLT is designed to transform XML into (a different) XML. XSLT is good to achieve interoperability between XML API's and the like. It's way overkill (complexity, speed, ugliness, difficulty).

XSLT is actually the opposite of what you want. What you want is a very simple interface to change the look and/or content of the website. Changing the content through XSLT parsing it's that big of a deal (Although, you'd need to write some caching code - because parsing on page load isn't acceptable).

But consider, you need to impose a very strict "format" for the page. So each source xml page would look something like:

<page name="page.html">
<title>This is the page title</title>
<message>
this is the message
</message>
</page>


That's all fine and true. And allows RRM (or who ever) to modify the content with ease. However, what if he wants the page to be different? Or wants to change the look of the site? Then you've got to go and modify the XSLT (which is a full blown turing complete programming language).


What you're looking for is a templating engine. A decent PHP choice is Smarty (http://www.smarty.net/)

It give flexibility, compete separation of programming and design. A disadvantage is that HTML is put throughout the template. So if the maintainer is uncomfortable with it, it is a lot more difficult.
But I think a better way would be to convert the site to a simple PHP content management system, like Wordpress. However, Wordpress does take more effort on the webmaster's part in maintaining, as upgrading webapps is a pain. And the webmaster would need to be experienced with PHP, MySQL, diff+patch (preferably) and perhaps some more.
Agree with you on all points. This has the advantage of the solution I outlined about (templating engine) of the owner not having to have HTML in his face (the CMS will separate that from the design).

As for the hosting. We don't need to change server, I guess. And if we do, I suggest we move over to Nearly Free Speech. Any tech lover loves it :)
While I've never used them, I love their philosophy.

But can we have an overview of which sites belong to "Wai"? And could we perhaps have access to the source files? Then we can begin planning some changes.
Good question. And about the "source", with the exception of this forum, it all looks like static content that could be scraped easily.
halfgaar
Posts: 150
Joined: Sat 23 Feb 2008 19:48
Location: The Netherlands

Postby halfgaar » Mon 11 Aug 2008 00:46

I don't want to spend to much time arguing XSLT, but let me say a few more things. As opposed to some kind of custom page definition files, I take XHTML 1.1 as input (because XHTML is simply XML), and have only a couple of lines XSLT to add header and footer info, and output HTML 4.01 with xsltproc's HTML 4.01 output writer. Easy as proverbial pie.

And because it's all XML (something which Smarty is not...), you can do very funky stuff with it, like generating chapter numbers and generating lists of references based on all the anchors and links in the page. I even was able to include the RSS feeds of some blogs as HTML code with only a few lines of XSLT.

And about Smarty; might as well use plain PHP (article written by a friend of mine).

But anyway, I think we'll be using a PHP CMS, so this is not really an issue.
Good question. And about the "source", with the exception of this forum, it all looks like static content that could be scraped easily.
Mostly, yes. But you never know what files lurk on the hard drive of the webmaster that are not to be found on the webhost.

To begin, that overview of all domains would be nice; there are a lot of them on which the Wai site can be found, and some of them contain information that is mostly the same, but not completely. This makes unifying it to one domain somewhat difficult, because different versions of articles can exist; a situation that should be remedied.
Havas
Posts: 51
Joined: Mon 14 Jul 2008 03:59

Postby Havas » Mon 11 Aug 2008 05:34

halfgaar wrote:I don't want to spend to much time arguing XSLT, but let me say a few more things. As opposed to some kind of custom page definition files, I take XHTML 1.1 as input (because XHTML is simply XML), and have only a couple of lines XSLT to add header and footer info,
Ok, use a very complicated programming language to include a template. It's your prerogative, but you shouldn't suggest it for someone that doesn't have an extensive knowledge in it.

Seriously, have a look at a 3 page example: http://en.wikipedia.org/wiki/XSLT#XSLT_examples


Now I'm going to achieve the same thing, templated:

Code: Select all

<?xml version="1.0" encoding="UTF-8"?>
<html xmlns="http://www.w3.org/1999/xhtml">
<head> <title>Testing XML Example</title> </head>
<body>
        <h1>Persons</h1>
        <ul>
{foreach $person}
           <li>{$person[name]}</li>
{/foreach}
        </ul>
</body>
</html>
And to be achieve the same amount of flexibility that mine has, you need to edit 3 separate pages for XSLT. So honestly, what do you think is better?

(And don't say my example isn't fair. I didn't include the style sheet - otherwise it would be 2 pages, but yours would be four pages. And don't say I didn't include the PHP processing code because it's irrelevant (and the example for your didn't include the server side code either).


Anyway, I have nothing against XSLT. I've used it to transform raw stock prices to (statistically) processed data. And it's useful for that. But to use it as a template language is like teaching someone Japanese so that he can order some Sushi in the waiters native tongue.


:)
and output HTML 4.01 with xsltproc's HTML 4.01 output writer. Easy as proverbial pie.
Nice. Turn XML to a format less strict (SGML is it?). Cool, I guess. Any particular reason?

And because it's all XML (something which Smarty is not...),
Sure. But that's because it doesn't need to be. But if you want it. The programmer needs to literally add two lines.

$smarty->left_deliminator = '<'!--;
$smarty->right_deliminator = '-->';

Well, it's valid XML now. (and if you like pain, but want more information available to your parser you could instead use:


$smarty->left_deliminator = '<';
$smarty->right_deliminator = '>';

..but don't say I didn't warn you.
you can do very funky stuff with it, like generating chapter numbers and generating lists of references based on all the anchors and links in the page. I even was able to include the RSS feeds of some blogs as HTML code with only a few lines of XSLT.
Sure. You can also make a 3D openGL game in it. Just cause you can, doesn't mean you should. Plus again, the complexity just isn't worth it. And it's MUCH easier to do in smarty than XSLT anyway :P
And about Smarty; might as well use plain PHP (article written by a friend of mine).
He makes a semi-decent point, but it's more emotional than anything. Smarty is far easier for someone with HTML knowledge to work with than PHP. It's safer (You don't need to worry about them doing anything malicious, or accidentally introducing a security vulnerability etc.) and it's not cached. In smarty the entire site would be cached.


And that's a big deal. I can give you write access to my template folder (ensuring no execute permissions of course). And there worst thing you could possibly do is vandalize my template, which I would shortly restore. Give the person access to a server-side programming language. They can completely mess up your machine, on purpose or accidentally.
But anyway, I think we'll be using a PHP CMS, so this is not really an issue.
It's not a bad idea. Also, you do realize that any CMS will use a template engine like smarty? For instance this board (phpBB) uses one extremely similiar.

It just makes sense. You'll have trouble finding a website that uses XSLT for template's, but prove me wrong :)
halfgaar
Posts: 150
Joined: Sat 23 Feb 2008 19:48
Location: The Netherlands

Postby halfgaar » Mon 11 Aug 2008 13:17

About the XSLT, let's just say my use case is a lot different than you're describing. I'd love to discuss more about it, but it's kind of off topic in this thread.

RRM, how about the working directories :)?
johndela1
Posts: 968
Joined: Fri 31 Mar 2006 03:54
Location: Portland, OR
Contact:

Postby johndela1 » Tue 12 Aug 2008 01:24

Havas wrote: Halfgaar, I'm not sure why you'd want to use Ruby? It's not standard, its slow as molasses(PHP is 3 - 9 times the speed: http://shootout.alioth.debian.org/gp4/b ... &lang2=php)..
This is a bit off topic...

the biggest resource is developer time not CPU time

I recommend only optimizing software when you actually have measured it under a profiler and have found a bottle neck.

If the site or app is fast enough then it won't help to use a lower level language to make it faster it will only make it more time consuming to develop.

Higher level languages like Ruby (or Python) usually result in less time writing code and easier times maintaining/debugging.
halfgaar
Posts: 150
Joined: Sat 23 Feb 2008 19:48
Location: The Netherlands

Postby halfgaar » Tue 12 Aug 2008 09:06

The slowness of Ruby wouldn't have been an issue the way I had it in mind; it would merely have been used for offline processing. The webmaster changes something, runs Rake (Ruby version of 'Make'), and all html pages are generated. The webhost itself would serve only static pages, so no PHP processing would be necessary.
User avatar
RRM
Administrator
Posts: 8153
Joined: Sat 16 Jul 2005 00:01
Contact:

Postby RRM » Wed 13 Aug 2008 19:41

halfgaar wrote:RRM, how about the working directories :)?
Huh, what do you mean?

Who is online

Users browsing this forum: No registered users and 2 guests