BurningbirdRecovered from the Wayback Machine.
I threw caution to the wind and submitted my carefully crafted session proposal to O’Reilly for ETech. I thought about posting it here, but is that bad luck, or poor taste?
Regardless, I will tease you and tell you that is is very complementary to what Sir Tim really wants but doesn’t know it yet. He looks for a revolution and Great Things and a crescendo of meaning; the rest of us just want to find things.
Silly things.
I am currently working with a small company to create an online store. When finished, I’ll point you in its direction–one of those very rare times when I’ve worked on a site that actually has a public interface I can point people to. Go, me. Go, company.
One decision we made right from the start: you don’t code from scratch when working with a common functionality such as a shopping cart/store front–you use existing code. Among all the many available packages we reviewed, we decided on using an open source PHP/MySQL solution, OsCommerce. One big difference, though, on using it straight out of the box is that I’m cleaning up the publicly accessed pages so they either use the Smarty Template engine, or simple and easy to use function calls that pull in the appropriate data. OsCommerce currently embedd barely wrapped functional calls to the database directly in the public pages, making them, frankly, a real mess for anyone but the most proficient PHP developer.
Once I create the non-business specific wrapper, regardless of what approach I use, this layer will go into the public domain, as a contribution to the open source community. Should be a satisfying effort.
There are other tools built on OsCommerce we could use; in fact, several. But they’re either commercial products with too restrictive licenses, or just about as messy in the public pages as OsCommerce (by ‘messy’ I don’t mean bad; I mean that there isn’t enough separation of the presentation from the process and the process from the data).
In other work, I also have the Rodent Regatta port from WP to MT and from HTML tables to CSS almost done, except for that damn problem with the vertical sizing of a contained element that is floated. I know about using clear:both in an element as the last element in the container, but I’m doing something wrong, it’s not working correctly.
Anyone spot what I’m doing wrong, or what I need to add?
Finally, I’m working with a couple of other people on a different site called the IT Kitchen (no relation to Doc Searls IT Garage–unless he wants to hook up, and he and the garage would be welcome). This site is going to host a two week interactive clinic focused specifically at non-techs, explaining as much about all of this as possible. Not everyone who programs is a professional; and not every non-geek weblogger wants to have others handle their CSS and basic site maintenance.
It’s going to be using a combination of technologies to ensure an interactive element, as well as provide a little something different. Everything will be Creative Commons or GPL licensed, and the static portion of the clinic will get wrapped into zipped files for copying when finished; the wiki and other interactive elements will, hopefully, continue to thrive on. Sort of a Wikipedia for webloggers.
(More on IT Kitchen later this weekend. )
I’m looking for volunteers, geek and non-geek, to help with this. Something like this is only going to work if its community driven. And If I don’t get enough volunteers, I’m going to continue quoting existential philosophers. Many more existential philosophers.
Speaking of existential, I’m finishing up my proposal for O’Reilly’s Emerging Tech Conference. I’m rather fond of it, but the success of the proposal is going to depend on who is judging the entries, and what their current focus is.
I’m starting to get a pretty good handle on what WordPress 1.3 will be offering, both from the code and forum discussions. I also found a link in the support forum for a WordPress 1.3 release wiki, which details the individual changes.
(From the wiki software used, I wonder if the WordPress documentation wiki is being moved to MediaWiki – the same wiki software used for Wikipedia. We can only hope.)
I’ve discovered some of these changes with the work I did for Doug, am doing for Steve right this moment (you can see the actual transform taking place here, as I migrate the HTML table layout to pure CSS), Tim (if he doesn’t lose his heart to Tinderbox), and might be doing for Loren (if he has a mind for this direction, now that he knows I’m not disappearing at the end of the month).
One big change, and one I adore, is pagination. With this, search results and archives are now paginated to list twenty or so entries at a time, with navigation automatially handled to go back and forth in the list. If you’ve ever searched my site on say, ‘flower’, my but you’ll kill both your and my bandwidth. Pagination will eliminate this problem.
The developers are also providing a Dashboard, I think for an overall linking mechanism to the site and the features, but it’s still under development–right now there’s nothing in the page. and you’re redirected to new posts. They’ve also packaging themes into their own subdirectory, and you can install and switch between themes just by coping the files into the directory and clicking a button in the administration pages. This will be good if you like to play around with your site a lot.
Architecturally, for the tweakers, the individual global values that you used to access in code previously have been added to a general object, but the old global values are maintained for backwards compatibility, at least for one release. The organization of the pages has changed, with a new header and footer page, which should help make the pages a little cleaner to work with. Recently when I talked about keeping your weblog tool independent, I mentioned about a split between the content of the page before the posts, the post listings, and the content after the posts. This is mirrored in the WP page split between wp-header.php, index.php, and wp-footer.php. I like it.
There’s also been function changes, and I’m still exploring what these are. From what I read, existing functions will be supported in a deprecated state for at least one software release. Good for those who have tweaked their pages.
Multiple blog support isn’t there, but an interesting announcement was made last week for what sounds like might be a parallel branch of development called WordPress MU. This isn’t a ‘fork’ in the code (i.e. a new and separate development based on original code) as much as it is a ‘wrapper’ around WordPress, from what I’ve read. The announcement about it says that it provides both Smarty template support and multi-user/multi-weblog support–in hosted environments.
The multi-user/multi-weblog capability should make WordPress more attractive for those who need a classroom solution to weblog hosting – a tool that can be used to create many weblogs for many different people, but administered from one spot.
As for the use of the Smarty template system, I am curious if this can be dropped into a regular installation of WordPress, for those who would prefer Smarty over the embedded function calls. If so, this would make a nice option to those who are uncomfortable messing with PHP code directly. Where before a person did the following:
<?php the_date(‘’,'<h2>’,'</h2>’); ?>
They would, instead, use a Smarty template tag such as the following that would resolve to the function call:
<h2>{$smarty.now|date_format:”%Y-%m-%d”}</h2>
The sidebar is also split off into a separate file. This does concern me a bit from a design perspective as this tends to enforce a specific type of weblog look, the two column look.
As an aside from a discussion of WordPress 1.3, when Movable Type announced it was providing a dynamic PHP-based wrapper around Movable Type, I thought that the company would take core bits WordPress and modify these to point to their own database so that WordPress plugins and templates would work with the Movable Type database. With this, though users would have lost the Perl plugins for MT, they would have gained the PHP-based WordPress plugins. At least, those plugins that deal with parameterized data, only.
This wouldn’t be all that complicated either. Smarty could have been used to transform tags for the traditional MT users; while others could have used the WordPress embedded function calls (and themes) if they wished. Licensing wouldn’t have to be an issue because MT could continue to license the MT ‘engine’ with associate Perl code, and GPL’d the code for the PHP wrapper. They would gain friends from the open source community, while the supported, proprietary, corporate Perl code would still be there for corporate types who get nervous around the word ‘open’.
And wouldn’t this have been an interesting way to mix proprietary and GPL code?
Another new, or I should say heavily modified, feature I accidentally discovered with the RC 1.0 version of Firefox (I don’t think it was in the .9x releases) is the “Find in this Page” text search capability. Previously, the search function was a window that would open, you’d type in the phrase, and it would scroll to item.
Now when you click the link for “Find in this Page”, a bar opens at the bottom of the browser. As you type in the word you want to find, Firefox immediately scrolls to the first word that matches the letters as they are being typed. And it works very, very fast, too.
You can also highlight the words if you want to see all occurrances in the page. And leave the Find bar up for use in all your pages, if you’re doing some heavy researching on a term.
Firefox beats out any other browser I’ve used, on either my Mac or my Windows laptops. Unfortunately, I still have to use IE for printing on my Windows 2000 laptop because Firefox does not work well with my printer. In fact, I have to re-boot the machine to kill the runaway process triggered by Firefox if I forget and do a print. But the HP drivers I have for this printer are badly behaved anyway (I must see if there are updates that fix the problem) so I’m not blaming the browser.
Tripping over to Loren Webster’s In a Dark Time weblog, I was surprised to see “Domain expired”. However, it’s been renewed so more poetry should be forthcoming. If you’re thirsting for the words now, then you can use his “tilde” URL – http://gemini.hmdnsgroup.com/~loren/In_a_Dark_Time/. Now might also be a good time to go to his front page, because I can see he’s added new material. You won’t be able to comment, though, or access a permalink unti his DNS renewal has propagated throughout the threaded void. The registrars say up to 72 hours, and an expiration can take longer than a new domain. Still, less than a day, I estimate.
This is a good time to mention that several .info registrars, such as my own, Dotster are giving away free .info domains. One site started it – DomainSite, but they’ve since gone to a fee of 0.99. Which is still very cheap.
So far, I’ve snapped up the following domains:
burningbird.info
dynamicearth.info
lampsurvivalguide.info
netsurvivalguide.info
practicalrdf.info
missouriphotos.info
prettypics.info
shelleypowers.info
mirrorself.info
yasd.info
aboutwomen.info
ladygeek.info
ladygeeks.info
photobird.info
solarlily.info
itkitchen.info
bookofcolors.info
saltofthesea.info
reallysmartweb.info
reallysmartpeople.info
hintofspice.info
sensualist.info
pragmatist.info
poetryfinder.info
Some of these really beg for a site, don’t they?
There is a limit of 25 domains per registering entity, in this case myself. This may only be for each registrar, but I think 24 domains is enough, don’t you?
Why is this happening? Most likely to begin to fill up namespace for this new TLD (Top Level Domain), and start generating interest in the .info domain. This domain is an unrestricted domain, which means anyone can register a domain name–here’s your chance to have a free domain for one year. Next year, of course, it will cost you.
Going back a moment to Loren’s site, the use of the tilde (’~’) has been around since Apache was a pup, I believe. My first site was a ’tilde site’, before buying my first domain (not easy to obtain in the early days; free at first, and then horribly expensive until Internic decided to allow competition).
A tilde site is one in which multiple accounts are hosted in one shared environment, and each account has a specific name attached to it, such as my own, ’shelley’. These are hosted in a specific directory, usually labeled the ‘home’ directory. A person can then access the site using the IP address or general name for the server, followed by a /~(name), and the web server, at least Apache, will serve up the pages until whatever domain name they use propagates through the system.
So if I, in my gluttony and greed for domains, happen to forget to renew my burningbird.net domain registration (horrors!), you could still access my site using http://gemini.hmdnsgroup.com/~shelley/, or this weblog using http://gemini.hmdnsgroup.com/~shelley/weblog/.
Of course, using the tilde site to access the page could cause some interesting challenges if you’re using the top level relative URL, as I am using for some my of my stylesheet effort. The reason why is that the top level domain in this case is no longer burningbird.net, but gemini.hmdnsgroup.com. And there is no /look, /photos, or /mm subdirectory at this location.