My notes follow:

The WordPress Style

Architecture: Theme support in WordPress encourages, but doesn't require, both modularity and standardization. While only two files (style.css and index.php) are necessary to create a non-trivial theme variation, WP's display engine recognizes a dozen-and-a-half file names and uses them in appropriate places if they are present. This creates a powerful incentive for most designers to stick to the standard file set. Those who need (or prefer) to use a non-standard file set are accommodated by what I take to be simple PHP coding. (Recall that one of my purposes, here, is to learn PHP; I'm not there yet.)

Presentation: Mechanically, WordPress is like lots of presentation engines: It pulls content (including menues) from a database, which it wraps in (X)HTML using a combination of PHP, PHP extensions, HTML tags, and the flags (the strongest are usually but not always labeled "class") which give CSS the hooks required to control presentation. This is stuff I understood in principle when I started this little project; what I'm about, these days, is coming to grips with the details. The reason I mention it now is that I'm annoyed with the CSS that styles Gangway.

Style: In the past week I've called LightCMS "well-crafted" (Sreejith is a code artist), Cutline "workmanlike" (Chris Peterson's a problem solver), and ModernPaper "delightful" (Brian Gardner's unusually disciplined). Although they're very different in detail, all use the same basic CSS vocabulary for describing/organizing the document. Since I don't follow the CSS discussions, I don't know what standards someone's trying to enforce, but I've read enough code in my life to have preferences. CSS is a rather spare coding language, but you don't need to look at many stylesheets to learn that there are a variety of coding practices (normally I'd call these "styles," but that would be confusing), and that some of those practices are more readable than others. Gangway's style sheet fails the readability test.

I have two problems with Gangway's CSS: It doesn't follow what seem to be the WordPress community's labeling conventions, and it's "clever." I'm sure there are good reasons to call the sidebar "addbar," the header "branding," and the meta area "tools," but if this theme called the standard WP design elements by their common names it would remove a barrier to understanding the (GPL'd) code. The cleverness issue shows in the style sheet's arrangement; Gangway's coder likes to group things for efficiency. Since CSS display reflects sequencing, it's possible to code for performance by sacrificing readability; that's what I think is going on, here. Can't say I like it....

Despite my complaints, this is solid code. I've learned what I wanted about sidebar design, which was my main purpose; I can move on to something else. Ajax-Berlee, methinks. Ought to be an interesting encounter.

This is a sweet design. Although it doesn't look a whole lot like Cutline, both the page layout and the underlying architecture are quite similar. But this theme's the product of an excellent coder, Brian Gardner (also author of the very popular Revolution theme, I see). I am, of course, a bit concerned about the 930 pixel graphic in the header.

The page layout is similar to Cutline's: two navigation columns, both on the right, and a moderately wide (albeit disguised) graphic which defines the width of the header, and thus the page. The details differ quite a bit from Cutline's, and the layout's not as visually attractive, but this is an excellent content-presentation layout.

Behind the scenes, there's a slight surprise; Gardner's used the Home.php page the way most theme designers use Index.php, and Gardner's Index.php is used like most designers use Single.php. (That realization sent me to the Template Hierarchy, where I convinced myself that the design decision makes sense, though it's unconventional.) Opening the files to examine the code is delightful: Gardner writes clean, compact, and obvious code, uses XHTML tags as they're intended, and organizes things well. Needless to say, his CSS files are similarly impressive.

I like this one, folks. It's not cutting edge stuff, but it's extremely well-crafted. Reworking it to fit my brother's screen would be a bit of a chore, but the necessary changes are well within my abilities. But let's look at something else for a few days....

One "simple" solution for a WordPress theme is to build separate, but stylistically consistent, templates for the various types of WP pages--the index, home, single entries, what WP calls pages (I might call them essays), catalogues, etc. That's Cutline's approach, and it's a nice implementation. The theme looks good, is fairly easy to customize, supports catchy graphics, and loads quickly. These are virtues, and collectively they explain why this is among the most popular WP themes.

The code that makes this work isn't especially interesting. It's workmanlike, and (let us say) uninspiring. To this HTML codeslinger, some of the tricks are a bit unsettling; in particular, Cutline uses the <Hx> tags in ways which undermine any hope one might have that those tags, by implying structure, add meaning to the document. (That's not to say I've never done something similar; it's just an observation about the coding style.) The CSS is similar, as you'd expect; it works, but another coder would have solved some of the problems differently.

Then there's that 970 pixel graphic across the top. Looks terrific, but it runs off the edge of the screen....

A first report follows.

One danger with templating languages, whatever the environment, is the temptation to build something really elaborate. LightCMS avoids this. One flexible page template handles all the display variants--front page, search, single item, categories, whatever. The great virtue of this theme is its simplicity: There are just a handful of files, and the template's primary CSS file controls the screen layout by doing really obvious things in a structured, well-organized fashion. Understanding this one's simple; it's a good place to begin exploring WordPress, and could well be a useful foundation for a more elaborate structure. I might return to this one when I'm ready to implement my final design.

Or I might not. One problem is that the design's so simple--it's well-crafted, but it's not really interesting. Another is that the CSS template implements a layout that's so single-minded that significant design changes would require totally rebuilding the main CSS file. I played with that idea for a few hours last week, then decided it was time to move on.

Joan says: "Whatcha been doing all night?"
"Moving Dabbler's Journal," I replied.
"So you're going to start writing again?"

I've done this before....

Installing WordPress was pretty easy, just like they say. Getting to the point you can make the installation requires some technical expertise; it's not difficult stuff, but it helps a whole lot if you already know how to do those things. And, while displaying the themes which can be used to "skin" the site is remarkably easy, if you want to modify them you need still another set of technical skills.

Some notes about the installation experience. These comments are not intended to be critical; I'm just recording my reactions to things that happened. Nor should these be mistaken for instructions. Although a couple things went wrong over the course of the evening, they turned out to be frustrations, not fatal issues. One was self-inflicted.

Wednesday

Wednesday evening I told John Skilton, who's been hosting this domain for several years, that I planned to move the site to a different host. John's hosted my other website, A Fan's Guide to the Midwest League, for about a decade, and that site will remain at Skiltech. But I'd decided to resurrect this Journal in WordPress, and it seemed preferable to move to a host which was familiar with the software.

Thursday

I examined the hosts list on the WordPress website, and decided to go with Laughing Squid, largely because their website doesn't make me think of K-Mart. Signed up for an account, played with their administrator interface (Plesk) for a while just to get a feel for the tools, created the MySQL database WordPress needs more or less in accordance with the installation instructions, then set up/transferred my dabblersjournal email accounts. This is familiar stuff for me, with tools which are not so different from some I use elsewhere; other folks might find it intimidating. Had I stayed on Skiltech's servers, the MySQL load would have added some complication, and he'd likely have had to enable PHP for me, but we'd have worked it out.

The next step, actually moving the domain, seemed like it should be easy--except that Wild West Domains' website denied that they knew who I was, regardless of the information in WHOIS. After a couple largely-wasted hours, I figured out I needed to log in via Skiltech's reseller interface. Once I solved that, the domain move was pretty straightforward if you've done this sort of thing before. BTW: Last time I moved a domain it took several days to propagate the DNS changes; this time the traffic had almost completely moved to the new host within a couple hours. Things have improved.

On to the WordPress install: The instructions on the WordPress website are excellent. My first two tries failed, probably because I was attempting a (slightly) non-standard setup (see my note, below), but in the end I just went with a standard file arrangement. That install, friends, was easy by almost any test. This installation requires that you understand ZIP (or tar) and FTP; I'm an old BBS sysop, and remember those protocols' ancestors. Non-techie mileage will vary.

Then I fired up Plesk to do a little tweaking, as WordPress needed some read/write privileges changed. This may a skippable step; it likely depends how you intend to use the software.

Some final activities for the evening: Took a quick tour of the WordPress menues, changing a few defaults to meet my known prejudices. Selected one of the available themes (truly easy; this is a great implementation). And I created a couple sample entries just to see how things work, and look.

Friday

Friday I played with WP Themes. I had several objectives:

1. Identify two or three themes I can live with while I work toward a permanent solution.
2. Identify five or six other themes that can teach me things about design and feature possibilities.
3. Begin to unravel the interlinked mysteries of PHP, WP templates, and CSS.
   • I've worked with several templating languages over the years so I don't expect serious conceptual problems, and I've used CSS casually on several projects but don't really know it. PHP, on the other hand, is completely new to me and one of the reasons I'm taking this migration on.

I have some look & feel expectations, too: A good theme is light-weight (that is, quick to load), looks good on a narrow screen, is visually interesting, and has the basic navigation features I expect of a weblog. Although I'm certainly not opposed to color, the print in the content sections must be black (or near-black) on a white background. And I'll likely end up with two nav bars, both on the right side of the screen, as I strongly object to designs which run far off the bottom of the screen.

I've also some opinions on how the front page should look (and work), but I'm willing to implement those features myself after I master the environment.

I quickly settled on Cutline (Chris Pearson), Ajax Berlee (Nico Berlee), and LightCMS (Sreejith) as themes I'm willing to deploy while I'm figuring things out. Quick comments on each:

• Cutline is just gorgeous. There are many details I want to change, but there's lots and lots to like about this design. Narrow screens are a problem, though not hopelessly annoying, and there are a few functionality issues I need to understand better. And this theme has the best 404 page I've ever seen.
• Ajax Berlee's front page is the closest to the layout I want, though it displays atrociously on narrow screens. Nico's scripting skills seem, at first glance, to be pretty impressive; I need to look at that code. I think I don't like the way he handles fonts, though.
• LightCMS is a remarkably clean implementation, though the small fonts are painful; it does work well on a narrow monitor. I've already discovered an issue in the way the program displays nested categories (in IE6--works acceptably in my other browsers), which I'll need to address soon if I want to continue to use the theme. But I really like it.

Several others caught my eye: Evernight (Ricardo De Sousa) looks like a good starting point for further explorations, Kolorful (Myo Kyaw Htun) is a fine minimalist implementation, Superfly (MiloIIIIVII) is really pretty, and Ice (also MiloIIIIVII) seems nearly perfect, albeit perhaps a little fussy. I'll be playing with themes for a while, yet, but that's where I got to on Friday.

The themes are easy to change once they're installed on the server. Installation, though, requires competence (again) with ZIP (tar) and FTP clients, and a willingness to explore the WordPress directory tree. These are fairly common skills in some environments, but in this context "simple" clearly depends on what you already know.

Saturday

Ignoring a whole lot of non-weblog activities, I spent Saturday looking at the templates, and the scripts in the templates. Made a few, tentative, modifications....

My preliminary assessment: This project promises to be interesting. A different person, though, would be better off taking WordPress up on the free installation offer, or perhaps using the Wordpress.com hosting service.