It’s amazing how remote all things internet get when the internet is not at your fingertips. I moved to my home town in Russia for the time being, and now I don’t have the kind of access to the internet that I’ve became accustomed to. I am back to the internet as luxury from being used to having it as commodity. I am back to dialup at 44kbps, that’s what my laptop’s modem gives me.

Even though it’s dialup it’s not cheap, so I can’t spend as much time online as I would like. So I am changing my internet usage patterns. Downloading all email to my notebook from Yahoo! Mail and Gmail instead of reading it online. Writing emails offline, and then sending them all at once when I am online again.

Oh, and now blogs. Blogs are such pain in the ass, pardon my french. I simply don’t have the time online to read them, and that hurts a lot. I am using the R|Mail blog to email service now to get all blog entries in my email that is downloaded to my notebook, and then I read it all offline. Still, it’s not perfect at all. A blog entry in full is a rare species. Usually it’s a thing with a “more” link that you have to click on to read the complete entry. So the rest of the entry does not make it to my email. Oh well. I think you can feel my frustration now.

ADSL option is available where I stay, but it’s not cheap either. It’s $40-$45 for 64kbps link and 1Gb of incoming traffic, or $75 for 2Gb of incoming traffic. Outgoing traffic is not counted. You pay hefty installation fee and the cost of two ADSL modems, and in a little while you’re all set. Only it seems that I use about 10mb per hour while browsing or downloading my email on dialup, which gives me… about 100 hours of browsing before I run out of my prepaid ADSL traffic allotment. Forget about downloading something interesting.

I get nearly everything I used to read online in my mailbox now, with the exception of few forums that don’t have the RSS support. Another excepion is the wiki that I maintain. Often than not I have to go out and read the blog entries in full while online. That’s how I burn my dialup minutes at disturbing speeds. Everything moves fast on the internet, and I do need to know of happenings earlier, not later. Anyhow, it’s a very different setting from what I used to have in the States, but I’m getting used to this all. It will work out, eventually.

One thing I understand better now is why and how many people here don’t have that feel of the internet and its surroundings that inspires me and other people with new ideas and killer applications. They simply don’t feel the pulse of the internet.

I keep a lot of bookmarks around. I mean, a lot! That makes it difficult to manage them, and for that reason I became a fan of tags, and a user of del.icio.us, a popular social bookmarking service. Now it’s much easier for me to find my bookmarks, and revisit places I’d been to on the net. I am talking of my personal bookmarks only, not the social aspects of having them on del.icio.us.

There are lots of other social bookmarking services now, each with its unique twist. The thing with most of them is that none are up to the level of convenience that I’d like to have. One thing I would like to have is a better interface, another is the ability to file my bookmarks faster, possibly skipping the step of tagging, yet being able to find a particular bookmark when I need it. Now LookLater might be the answer to this. First I heard of it from the post Startup blog, day 1 on The Business of Software. The idea was to have the first beta release out in 12 days! If you scroll down the original post, you’ll find out that Doug Martin, the author of the idea, is someone who really has the vision and experience to make it happen. He set up the development blog so that others could watch his progress in reaching the goal.

The first beta release did not happen yet, but it is about to happen anytime now. In the meanwhile, Doug released a couple of pre-beta demos, and they were quite impressive. It had that rare quality of being pleasant to use when I tried. I called those little bookmark pieces that were created on my clicks the looklets. Yet a very important thing to me was that Doug really listened to the users, and came up with things than can make me switch to LookLater indeed!

Anyway, don’t just read about it. Get out, read the development blog, try the demos, touch it, and feel what it’s like. You might like it as much as I did.

As a puzzle collector, I am an active user of eBay. eBay is invaluable source of rare and difficult to find things, including puzzles. I am also a user of sniping tools as they, on one hand, allow me to place my bids while I am away from the computer, and on the other hand, help me to place the bids at the very last moment, thus leaving little chance for someone else to outbid me. One tool that I used before was JBidWatcher, but then I switched to Biet-O-Matic, for it’s a more lightweight, yet capable enough tool.

Until now, due to the eBay’s pricing scheme for the use of their API calls, the sniping tools I’ve been using employed the strategy of scraping the content of the site. Not only did not that conform to the eBay’s User Agreement, it also had the tendency to break once in a while due to changes in the pages layout that eBay made routinely. Starting today, eBay dropped all fees it charged for the use of the API (when you use their new unified scheme), as well as the fees for the membership in the Developers Program. The programs and tools developed by people still require certification, but it’s free now as well, see Member Benefits and Fees.

Anyone who has been using sniping tools knows the feeling when a bid did not go through due to the tool being broken because eBay changed the page layout. Hopefully with the use of the eBay API these tools can reach a higher level of realibility that I’ve been looking for. I also had an idea for an interesting tool that I called “Concise eBay” about a year ago. After careful calculation, and reading the eBay forums, I decided against it. The idea still stands, and I might even get back to it sometime now, who knows!

Sometimes I think that having less choice is better than having more choice. It may depend on a person, but it can drive me crazy when I have too vast selection, and I have the need to make a choice. That’s true for all kinds of mechandise that I buy. That’s why I always look for the reviews, recommendations, and for a forgiving return policy.

I wanted a wiki for my site, as I find wiki a wonderful tool for collaborative development of ideas. I needed to make a choice, so I googled for wiki comparision tables, and reviews. I found several of them, and at first I chose MediaWiki as the most well known wiki out there. Well, after I have downloaded it and have taken a quick look around, it ticked me that it was way too big than what I imagined it would be. After careful investigation of the same comparison tables I put my finger on WikkaWiki that, compared to lots of other lightweight wiki tools is still being developed.

Installation was a breeze, quick and easy. I put a full copy of the distribution in a separate directory, and linked to it from where I wanted my wiki to be. I wanted several of them in different places and separate, and that way I was able to make that happen with ease. Everything happens on a single installation page, and after you have filled it with all necessary data, it creates the database tables for you, and initializes your wiki. The configuration file is put in the local directory, and you are free to modify it if needed.

Nothing is perfect, as you can imagine. For my installation, I wanted the wiki to be private, and for that I needed it to not allow freely available registration. There was no option to do that, so I had to hack the code, remove the registration part, and move it to a special administration page inside the wiki where I could create new users myself. Also, the defaults are set in the configuration file to allow public access to the wiki pages, and I changed it to allow registered users only. The most interesting (and somewhat disturbing) thing was that WikkaWiki creates a cookie when the user logs in, and that cookie is not linked to specific wiki, but rather to the host. That is, when you have several wiki on your site, once you log in to one of them, you get logged in to all other wiki automatically! I will see if I can fix that as that’s not quite what I want.

I’ve chosen WordPress as my blog engine for a number of reasons. One of the main reasons to try it was the promise of one-minute install procedure. And in fact I did install it just under a minute, to my surprise! Having done that, I set up to find a suitable theme that I would have liked, and did not really find one. But hey, it’s just me! One thing I did not want to have for sure was using one of those common themes that are replicated on hundreds, if not thousands of WordPress powered sites, even though I am not a web designer at all.

Here are some of the annoyances that I encountered while setting up the blog using WordPress:

• It was very unclear how to set up your theme elsewhere other than in WordPress internal themes directory. Blog was not the only thing I was going to have on the site, and I did not want my content to go somewhere deep inside the WordPress installation. After lots of experimenting, and reading of the WordPress code I found a solution. I set up a theme directory in my website root, and had the necessary constants set to point to that directory. Here is the contents of my /index.php:

  <?php
          if (!defined('WP_USE_THEMES')) {
                  define('WP_USE_THEMES', true);
          }
          if (!defined('TEMPLATEPATH')) {
                  define('TEMPLATEPATH', $_SERVER['DOCUMENT_ROOT'] . '/MyThemeDirectory');
          }
          require('./wp/wp-blog-header.php');
  ?>
  

• Wordpress support is a form of community support done by volunteers in the WordPress support forums. Developers seem to be rarely set to investigatate the problems, and that leaves experienced users no real option other than fix the problem themselves. I had this nasty problem with comments RSS feeds that I was fighting for over two weeks, and nobody could help me out there. I finaly figured out what’s happening, and put a fix in place in my copy of the code. I also submitted the bug report. The way it looks now it’s probably going to be fixed for the WordPress 1.6 release.
• For some things extra I put together a WordPress plugin for my installation that has a set of the functions I want. It is interesting, that a third of those functions are reimplementations of the WordPress’ standard functions that produce links, but do not provide the title attributes for the links, and have no way of passing those to them. I like having it all consistent and with the title for each link I have. I’ve also rewritten the calendar to not include any future posts, give me a slightly different interface, and (again!) provide the titles to the links. I tried to use some advanced features, such as replace the names of the templates that WordPress is looking for, and add some of my own using the template_redirect action, but failed miserably. Using the include()/require()/require_once() statement did not work and produced errors in completely distant places of the code. Partial success was when I used load_template() function, but in the end of the day that did not work either, giving me a set of similar problems for some of the templates I tried to include for specific actions. I finally gave up, after I discovered that it’s a long-standing problem. Needless to say that the support forum people were unable to help me with the problem (sigh).

As I studied the code a lot, I have an opinion on it. It’s a mixed opinion. The main thing is that the code just works for all standard applications and settings. That’s why people use it in the first place. Customizing it to suit your specific needs might be difficult, however, depending on what you want to accomplish. Some nontrivial bugs can drive you crazy, and I did not find much support in that area. I might not have looked in the right places for it though. The overall feel of the code that I’ve got was that it is structured well, yet it’s somewhat mushy, not clean and lean, and not very consistent. Also, it came as a shock to me when I turned on full warnings, and the system spit out a hundred of warning about undefined variables each time a blog page loaded in the browser. Hopefully the code gets improved as it matures.