May 27th, 2008

SEO Misconceptions

Search Engine Optimization, a rather popular, yet misunderstood topic. SEO promises increased page views by appearing as close to the top of a search results page as possible. Naturally, this requires optimizing towards specific keywords related to the content of the site being optimized. There are many common sense things that can be done to increase pagerank, but there is a ton of bad advice going around. After reading several articles on SEO which offered poor advice one can advise against some poor choices.

Multiple (identical) links to the same content are not bad. People tend to confuse, multiple pages containing the same content, and multiple links (on separate pages) to a single page that contains content. Cross linking is beneficial, not only for SEO but also usability. In WordPress (since 2.3), there are three methods of organizing posts, archive by date, category or tag. In most cases using only one of these for organizing data is not wise. Tags and categories and work together, as visible on this blog.

Contrary to popular belief, WordPress only generates one permalink to a particular post. Sure all the different ways to get to a post still work if manually entered. The links WordPress generates, however, result in only one location for a page, independent of taxonomy (it is all in the permalink structure). Need proof? Look at the link to this post. Visit the archives of the tags this post belongs to. Then visit the archives for the category this post is in. Lastly visit the by date archives for the day, month, or year this post was published. Notice how all the links are the same? This is by design, WordPress does this naturally, and it does not adversely affect pagerank.

Do not use nofollows. Simply put, the nofollow attribute produces a contradiction of intent. What is the point of creating an hyperlink if the user (be it human or robot) is not supposed to follow it? Sure, they do have a use for linking to outgoing content in things like user comment. However, in sections where the content is fully controlled (e.g., blog posts) why instruct the user not to follow the provided links. Under no circumstances should an internal link contain the nofollow attribute. If internal content is not trustable, then there are greater problems to worry about than pagerank.

Never substitute usability with SEO. A user may find a your site faster, but if they can not navigate the site they will not stay long. Someone suggested that a usability enhancer (breadcrumbs) should only be used on pages (not categories) as a SEO tip. Sadly, this is completely wrong and is detrimental to usability. Breadcrumbs should be placed in the same location on every page of a site if implemented. The single exception is for the front page, but thats an entirely different story.

-John Havlik

[end of transmission, stay tuned]

May 21st, 2008

Breadcrumb NavXT 2.1.2

Posted by John Havlik in WordPress > Plugins

The second maintenance release for the 2.1 series, and this month’s release brings some useful bug fixes. Static front page support has been improved (serious testing of it’s functionality took place in the past month.). This includes automagical detection of static front pages, which will do all the heavy lifting for you. Paged items now work as they did in 2.0. Do note that paged item support is not complete (that’s getting thrown into the July 2.2.0 release). Additionally, the behavior of the link current item option has changed, now instead of trying to actually generate a link the href is left blank. This seems to be the best way to link to the current page, as it is the simplest method for implementation and is guarenteed to work 100% of the time. Of course, that may change again in the future.

-John Havlik

[end of transmission, stay tuned]

May 7th, 2008

Some Fixes

Posted by John Havlik in WordPress > Plugins

When static front pages were reimplemented in Breadcrumb NavXT 2.1, it seems some things slipped through the cracks. In particular there is a problem with the frontpage breadcrumb, when on the frontpage. This and the options for static front pages were never cleaned up (my bad, I know). Another, relating to a /blog/blog and // error in the hyperlink for some by date archives was reported. Both of these will be fixed in this month’s service release, 2.1.2. For 2.2 support for WPMU will be improved (the administrative interface has problems with it right now), along with a widget. That release is a few months out still.

-John Havlik

[end of transmission, stay tuned]

May 2nd, 2008

Cran-Berry Build 100

Posted by John Havlik in WordPress > Themes

A little late, but fully functional. School work prevented the intended release date for Cran-Berry. Several styling, and code fixes were implemented which should provide a more uniform experience for most browsers. This includes new code for centering the content on the page, and some other general fixes. This release is not a full ‘stable’ release like a one for Breadcrumb NavXT, but this should be usable for advanced users. In the zip is two folders, one containing the theme, the other containing the modified version of WP-PageNavi. The other two recommended plug-ins should be downloaded from their respective home pages. Naturally, Berry implements support for these plug-ins in a safe manner–not having them will not cause PHP errors or warnings. Upon activation of the plug-ins they will display immediately (in most cases).

To change options, look in the functions.php and there will be a $berry_opts array of options a la Breadcrumb NavXT. These are documented in the source right above the corresponding option. Additionally, the sidebar is WordPress Widget aware.

-John Havlik

[end of transmission, stay tuned]

April 25th, 2008

Development

Posted by John Havlik in WordPress

Ok, so later today there will be a maintenance release of Breadcrumb NavXT, this will be version 2.1.1. The plan right now is to release a maintenance release each month with bug fixes and such. If enough new features are requested, a full version release will come out, and regardless a new version will always occur within a week of a new WordPress release. This release fixes some issues with the link current page option generating invalid HTML and some other bugs. Nothing in particular is slated for 2.1.2 yet, but it is reserved for any new bugs found in 2.1.1. 2.2.0 will add a widget for breadcrumbs to use with WordPress’ widget system.

Now onto Berry. Cranberry will be released sometime next week, preferably before May 1st. Currently, some tweaking to the styling needs to be done, especially for header images (logos), the search bar, and comments. Comments in general are getting overhauled. Right now this blog is still using the K2 comments.php, the only part that really isn’t my own code. This will be replaced by a new comments.php sometime this weekend. Then some hard core optimization needs to take place to speed up page load time and such. This first release will not have an Administrative interface for implementing modifications to it. The next release will contain options to change various things, centering around the amount of ‘flare’ the design has. Only one JavaScript line of code is in use right now, in the search bar. In the future more may be added and always these additions will be optional. Internet Explorer 8 still doesn’t play well with the theme, still need to figure out a centering method that it likes.

Lastly, some progress on WP Trainer may occur this weekend. The most likely thing that will get done is database support and some basic administrative interface. Since GPX files tend to get very large, often over 2mb, PHP will not want to parse them correctly, thus an external command line based tool will be included (with source) to split up activities in the files so that they can be uploaded. The last order of business that needs to be solved is a problem with Google maps not liking when too many points are placed on a map. I’m still not sure how to fix this but I don’t think it’s something a little searching can’t fix. It appears the common approach is to drop many of the track points, which is not always acceptable, so smart dropping algorithms may need to be implemented. Currently, the server requirements for WP Trainer are shaping up as so:

PHP5
- XML parsing support
- GD library support
Ability for PHP to write to files in a directory

I’m still debating if it is beneficial to dump the parsed GPX data into a database or not. It seems that it could dramatically increase database load and size, which is not always a good idea. Naturally, caching of data will be very important. Ideally, once a course is defined (named and has a track point map), any future runs on that course should automatically identify as using that course. This is very much a plaything project, which will take the backseat to all other paying or school projects. Development could always be expedited by sponsorship by a generous individual or organization.

-John Havlik

[end of transmission, stay tuned]