Inside the Soul of the Web
Google records and occassionally compiles a compilation of the searches that it runs. I'm not sure what it tells us--nothing good probably--but here's the end-of-2003 version.

As I recall, one of the older search engines (McKinley as I recall) let you monitor the most recent searches as they came in. They were typically related to the same sort of salacious topics as take up much of Google's CPU cycles today. Apparently there are screens showing a similarly real-time display at Google headquarters today.

Media Servers and Standards
A project in these here parts of late has been experimenting with media servers of both the music and video and the pre-assembled and homebrew varieties. As I accumulate more experiences in this area I'll be creating a separate archive that concatenates the various posts. However, before delving into hardware and software specifics, I'd like to make an observation about one particular annoyance/limitation I've run into when trying out the multitude of different products out there. That's the lack of standards.

I'm not talking so much about the formats for individual files. Although that's a bit of a problem too. Among "lossy" formats, there's MP3 of course. But there's also Windows WMA and Open Source OGG VORBIS to name just the most common ones. Both technical and philosophical debates abound surrounding all the formats. Then there are the "lossless" formats that are theoretically better for original source ripping and encoding--but disk sizes are still not quite at the point that they're suitable for general use because they take up more space than even high bitrate encoding using lossy formats. But, as I say, the fact that not all music servers support all formats (especially outside of WMA and MP3) isn't the focus of this particular discussion.

Rather, my concern here surrounds playlists--collections of files either manually or automatically (based on criteria such as genre, rating, or bitrate) selected from a collection. It can be laborious indeed to create playlists from multi-thousand file media libraries. But it's quite necessary to do so--especially when using the limited LED or LCD interfaces of many of the newer "appliance-like" players such as Tivo's Home Media Option (HMO) or Turtle Beach's Auditron. Yet, it's often difficult to impossible to transfer playlists among devices because of the lack of common standards. There are a few examples of cross-device support such as Tivo HMO's ability to stream files from systems running J. River's excellent Media Jukebox as well as its own software. And there are a few efforts underway to try and promulgate common open standards such as this one. But in the main it's a real pain to add a new photo or image streaming appliance to your network because so much work typically is required to manually recreate playlists or photo shows.

I've been on a bit of a slide scanning jag recently between getting photos from a recent trip to California up on my website, digitizing some old family slides, and getting ready to update the sea kayaking website that I maintain. For hardware, I have a Nikon Coolscan IV ED 35mm film scanner. Of course, Nikon supplies its own scanning software and there are any number of very sophisticated scanning software products out there. But, for my money, the best bet is VueScan from Hamrick Software. It differs from most of the other programs out there in that it's well designed for essentially batch processing a stack of slides that will later be image-processed using PhotoShop or the photo editing program of your choice. This isn't to say that VueScan doesn't have a considerable number of options. It does. In fact, you'll need to spend some time fiddling to figure out the settings that work best for you from multiple scanning passes to color balance to scan resolution. But these are settings that you'll often leave the same from picture to picture--or at least change only occasionally. However, VueScan doesn't have the sort of histogram and color curve manipulation features found in something like Nikon's own software. Which is just fine. I don't usually want to do a lot of image fiddling when I'm scanning a slide. I just want to get a decent scan onto the computer with the minimum of muss and fuss and do the fine-tuning with PhotoShop, which is far more sophisticated than any of the scanning programs.

There's a trial version available. It's well worth checking out, especially if you find yourself frequently scanning a stack of slides at one time.

I may perhaps be the last person to relaize the wonders of RSS (Really Simple Syndication). Essentially it's a means of obtaining a sort of headline feed from news sites and blogs. From a technical standpoint, it's a dialect of XML. From a user standpoint, it's the latest fussilade in the long-running push vs. pull battle. I won't go into all the details here, but the bottom line is that an RSS client can aggregate the headlines from all the blogs (like volokh.com) and news feeds (like CNET news) that you like to follow on a regular basis in a single location. It therefore lets you avoid having to either remember to individually check a bunch of sites or to signup to receive a bunch of individual newsletters. Many sites have their own RSS feeds; others are syndicated through third parties. For example, see syndic8.com.

As is often the case with the Internet, things with RSS are changing rapidly. From my experience, there's no ideal client out there. Browser-based plug-ins like Newsmonster are theoretically the best approach because the site contents are typically intended to be displayed within a browser anyway. So why not start off there? But in my experience the browser add-ons available to date just don't work--causing crashes and other sorts of failures. My favorite free Windows client is SharpReader. One nice aspect is that it lets you import and export lists of feeds. This is a workable alternative to server-based lists if you want to access the same set of feeds from several different PCs. Awasu is another not-bad alternative as is RSS Reader.

Website peeves of the week. Perhaps my brokerage house has a legitimate interest in ensuring that I'm using unique and difficult/impossible to guess usernames and passwords, but the same can scarcely be said of basic reservation or newsletter sites. Yet, you still run into the occasional website that rejects the "standard" username/password that I use for non-critical stuff, insisting upon more letters or random number permutations. Of course, I won't remember the oddball combination required--unless I write it down somewhere. Sure, keep me from using a username/password of "abc"/"abc", but insisting on 8+ character usernames and non-alpha passwords for casual site logons is a bit much.

The week's winner though is Amtrak. You'd think that they could afford a decent site design, but they've got significant problems. Want to change a reservation? Can't do it online -- you have to call it in person over the phone. Enter a discount number (such as AAA) which isn't available for a particular train (such as the Acela) and it will simply generate an error rather than continue with the reservation process while informing you that there's no AAA discount for this train. And, oh, when it bounces you back it wipes out all the credit card information that you entered so you have to start the whole process over again. You'd think at this point in the Internet's evolution, an Amtrak could come up with a decently designed e-Commerce site.