The Austringer

I have a Gateway MT6458 laptop. It came with Windows Vista pre-installed. I pretty much immediately shrank the partition to make room for a Unix boot partition and a shared data partition. I chose the Xubuntu Linux distribution to install for my Unix desktop.

I’ve been through one bungled upgrade that was my fault, and had to reinstall from scratch. Today, I tried upgrading to the latest Xubuntu release, 8.04 LTS “Hardy Heron”. And this time, it appears that all has gone well. I did have to change repositories, for a prerequisite noted on upgrading is that all recommended system upgrades be applied before the OS version upgrade is attempt. The “update-manager” application handles this, and it worked for all but two files: an update of itself and its GUI. I got a “Failed to fetch” error message. After retrying the downloads periodically, I figured that the issue wasn’t temporary. A bit of Googling revealed that sometimes better luck is had if one changes the repository used to obtain files. OK, fine, I thought. I was going to pick one essentially at random when I noticed a “Choose Best Server” button. That sounded like just the thing, so I hit the button. It took about five minutes, and it selected a host server in the USA on the west coast. When I tried the updates again, this time they went through.

Then it was time to upgrade the OS itself. I turned off all my running applications, then started the upgrade to 8.04 LTS. That took about 43 minutes to download everything needed, and then another hour or so to install things. Every once in a while, I’d notice that there was a prompt for my attention. After dealing with those, the process finished up and requested a reboot. I did that, and the system went through GRUB and I selected my Linux partition to boot to, and “Hardy Heron” did indeed come up without a hitch. So far, so good.

Warning buoys for right whales installed along Massachusetts Bay

Chris Clarke and the Ornithology Lab at Cornell University have an application for bioacoustics: right whale detection in shipping lanes. Right whales often make contact calls, called “up-calls”, and a series of ten deployed buoys with hydrophones and communications gear can pick up these calls for right whales within five miles of a listening buoy. Onboard processing does a first pass at picking out a “top ten” list of possible right whale calls, and those are uploaded to the Cornell Ornithology Lab for further processing. The system is computer-assisted rather than computer-automated, meaning that the computer processing narrows the things that would require a human decision, but it relies upon humans to make a final determination of whether a right whale call was present. If that is the case, the buoy is marked as having one or more right whales in the vicinity, and is tagged as having an “alert” status. This is reflected on a website that ship captains can access and, hopefully, reduce their speed while traversing areas where right whales have been detected. Right whales move slowly, travel near the surface, and ship strikes remain a major source of mortality for right whales. By highlighting where right whales are, Clarke hopes that responsible captains will take steps to reduce ship speed and post lookouts.

Abbie Smith is the latest blogger to switch over to ScienceBlogs, taking her ERV blog there.

Of course, there’s a tale there. ERV was formerly hosted on “Blogger.com”. Some time ago, Blogger was snapped up by Google. It’s still a large, free blog hosting site. Abbie was already planning to switch over to Seed’s ScienceBlogs, but was planning on doing so at the end of the semester so as to have some free time to devote to handling the inevitable little problems that crop up.

The schedule got advanced because last week “ERV” at Blogger suddenly disappeared from the Google search engine, and this morning the blog disappeared entirely, with a cryptic message as to how the weblog had been deleted and its name was no longer available for use.

Mark Chu-Carroll of the “Good Math, Bad Math” blog looked into the situation, found out that somehow Blogger’s administration had come to a determination that the “ERV” blog there was supposedly webspam, and vouched for Abbie’s good character. The “ERV” blog at Blogger is now back up.

Abbie is rightly indignant, though, that someone else would have to intercede on her behalf. She was given no notification, no opportunity to rebut a charge that her blog was webspam, no opportunity to collect her data, and — most importantly — no apparent way to contact anyone about the problem.

As I’ve gone over before, Google does have to balance the existence of web security problems against the information that it reveals to legitimate webmasters. However, in my opinion they’ve gone way too far toward the paranoid extreme of choking off communication with people running websites. Maybe it is arguably the correct thing to do with third parties who simply rely upon Google’s ubiquitous search engine to direct people to their sites, and who could be bad guys. I’d argue that there is still some need there for a bit more outreach. But in the case of someone like Abbie Smith, who is using a service provided by Google, there is more responsibility inherent in the relationship than simply sometimes providing a weblog host, and arbitrarily abrogating that relationship without communication and without means of arbitrating the outcome of those decisions. Google needs to figure out if they have what it takes to provide a service to bloggers, and get out if they decide that’s not the role that they want. It is not ethical to use the content generated by bloggers to bolster their bottom line without at least providing some minimal level of interactivity between the blogger and those who administer the blogger’s web presence.

I’m somewhere near Dade City, and I just got email that Diane’s weblog had been cracked. So I just upgraded Wordpress there and figured I might as well upgrade it here.

The upgrade worked without a hitch there. Here, I had a hitch. My customized theme doesn’t play nice with the new Wordpress, and my battery is just about dead, so until I get back to somewhere with 110 volt power, it’s just going to be using the default theme.

Picture banner and other features to return later.

Update: Apparently I hadn’t gotten around to copying the theme, which goes a long way to explaining why Wordpress wouldn’t work with it. There was a further bit with most of the sidebar disappearing due to the “sideblog” plugin; I’ve removed sideblogging, which I hadn’t been using lately, and it looks like the rest of the theme is doing OK.

Update 2: OK, I didn’t notice that my pre-2.5 widgets didn’t transfer to 2.5. That may have to wait a bit.

Amazon has been looking for more ways to market content, and they’ve come up with one. Amazon’s Kindle is a device that provides e-book content, but goes one step further: instead of cartridges or downloads to your PC that have to be transferred to your Kindle device, Kindle comes with its own wireless connectivity. This means that you can order and receive content for Kindle right from Kindle. That’s based on EVDO, so pretty much anywhere you get a digital cell phone signal is Kindle-ready. Here’s something they got right in this: the wireless access back to Amazon’s Kindle store is not billed, nor is browsing Wikipedia. Now, they do want about $400 for the Kindle reader and $10 each for downloads of New York Times bestsellers, so certainly they don’t plan on losing any money in this proposition.

Speaking of not losing money, the following Amazon Associate link will not only provide you with a way of ordering your own Kindle, it also nets me a 10% referral fee. So if just ten of you buy one via my link, I will be able to afford one of my own. (Currently it is sold with free two-day shipping.)

The Amazon Kindle page has various video endorsements and a 30-inch drop test to watch, as well as listing various books, newspapers, and weblogs that can be delivered to Kindle. The user reviews mostly are in the positive, but you’ll probably want to have a look through several before making a purchase decision.

I’m using a MacBook Pro laptop as my primary development platform for work. The project has sprung for the OS upgrade to the new-ish release of Mac OS X, version 10.5.1, or “Leopard”.

So today I identified a hard disk I could erase, plugged it into an external USB case, hooked that up to the laptop, and cleared enough free space to accommodate the upgrade process. I checked “Software Update” and got the 10.4 install up to date. After a reboot, I started the upgrade. It restarted again and began the process. I left the whole thing on its own for a couple of hours. When I came back to it, I had a login prompt. It stepped through a wizard-style process for registering the upgrade, then rebooted one more time. At that point, I was back to usual laptop function, but now with Leopard running the show. Xcode is OS version specific, so that had to be installed from the DVD, too. There were optional updates for the standard Mac applications, so I installed those, too.

All in all, the indications are that it was a pretty smooth in-place upgrade.

I’ll have to see have my experience holds up following that. I certainly wouldn’t do that with Windows; whenever I do a Windows install, I take things down to a new partitioning and clean install.

I followed up on a comment left in the earlier comment thread about giving gnuplot a try. For some reason, I had difficulties getting it installed on the MacBook Pro, but I got past that hurdle and spent some time today working with gnuplot.

OK, for someone who has access to Matlab, what exactly is the attraction to command-line plotting tools like gri and gnuplot? It is precisely because these can be called as part of a batch process that I find them useful. In doing the work I do these days, I have replicates of runs. If I use an interactive tool, I’m going to be spending a lot of my time simply stepping through a process over and over. If I put in some effort up front, I can write myself some scripts to automate the process and let me spend more time in analysis or coding.

“gri” does permit the production of contour graphs pretty easily, but I was finding it a bit rough in trying to get a surface rendered instead. That’s kind of frustrating, because there are examples online that look like they are doing what I want to do. So I thought I would check to see whether it was any easier in “gnuplot”. As it turned out, it was.

gnuplot provides two basic graphing commands, “plot” and “splot”. “splot” is the feature of interest for me, and in gnuplot 4.0 and above, a further useful style of “pm3d” modifies that. I have to admit it took a while for me in playing around with gnuplot interactively to figure out exactly what had to be done. One of the big steps was putting the data in just the right format for gnuplot to handle. I could dump my xyz data into gri, and gri would figure it out for gridding. Not so with gnuplot. The xy coordinates have to be completely sorted. Further, gnuplot expects a blank line in between each column’s worth of xyz entries. Fortunately, there was a site online offering an awk script for adding those blank lines. A little extra work in my Perl script formatted x and y values with leading zeros so passing the file through “sort” gets it set for awk to do its thing.

Then there were the data points, irritatingly plotted with “+” symbols. It wasn’t at all clear to me that what I wanted in getting my surface plotted without distraction was to tell gnuplot to “unset surface”. The “pm3d” style handles the rest if I use “set pm3d at s”, which tells gnuplot that I want the color mapping of “pm3d” applied only at the surface. Other ways to specify pm3d action includes “bottom” and “top”. For myself, the “surface” option does fine on its own. Another thing I settled upon was using “set hidden3d”, which obscures parts of the projection that lie behind others.

gnuplot offers a large selection of “terminal” types, essentially how one wants the output to be presented. On Mac OS X, the “aqua” terminal is the default, putting a plot onscreen. For my batch processing, I can just “set term png” and I have output suited for figures in LaTeX via “pdflatex”.

I haven’t worked out all the labeling, but that will come soon enough, I think. Here’s a code snippet from my Perl script to output a gnuplot command file and launching gnuplot on it.

open(GPC,">${dataname}.gpc");
print GPC < < "EGPC";
#set term post
set pm3d at s
set pm3d scansautomatic flush begin noftriangles implicit corners2color mean
set palette positive nops_allcF maxcolors 0 gamma 1.5 color model RGB
set palette rgbformulae 7, 5, 15
set colorbox default
set term png transparent xffffff
set out "${resid}.gp.png"
unset surface
set hidden3d
splot '${resid}.grid.gp' w d
EGPC
close(GPC);
        # Call gnuplot
        $cmd = "gnuplot ${resid}.grid.gpc";
        system($cmd);

I make use of the Perl “here” document format, which makes it pretty simple to get Perl variable values into a block of text from another source. The funky command parameters were obtained by saving an interactive command session and using that to put together the script above; gnuplot came up with most of that.

Here’s an example output graph:

Obviously, transparency is better in some contexts than others. But basically I now have an alternative to firing up Matlab and interactively generating a color-mapped surface plot.

I have a FreeBSD-based box acting as a file server. One thing digital photography does is allow one to accumulate a lot of photo files.

The system that seems to work for Diane and I is to put photos into directories named for the date and broad subject the pictures were taken, say “2008_203-hawks”. But the files are large, so browsing photos across the network can try one’s patience.

I started a script on Sunday that is working backwards through these photo directories. In each directory, it finds original JPEG files and applies the “jhead” autorotate procedure to them. If no thumbnails and other size photos are there, it creates an “alt” directory, then makes a thumbnail, a version sized for the web, and another version sized for use in PowerPoint presentations. The original, thumbnail, and web size versions are being put in the “alt” subdirectory, leaving just the PowerPoint sized files.

Those are large enough to provide a good view on screen, while being small enough to transfer efficiently across the home network.

The script has gotten back to photos taken in early 2006 so far.

Via FreeBSD, I’m using the XnView program for browsing. So far, that seems to be working well enough.

How does one plot 5.2 million XY data points?

I ran into this while working on a paper submission. This is one thing one does not lack for when doing evolutionary computation: size of data sets.

Matlab seems to become dog slow and unstable when trying to plot large numbers of data points. The interface bogs down such that trying to re-label axes is a real chore.

I tried out the GNU R package, and had it crash on trying to read in the data set.

Then I started going through plotting packages in the FreeBSD ports system. That’s where I came across the GRI package. This is an open source, GPL licensed graph plotting language. It has simple examples available online. As its documentation notes, it is a package with a fairly shallow learning curve. Its interface is entirely command-line, and its output option is PostScript. In interaction, it simply outputs PostScript graphic after Postscript graphic, simply named “gri-nn.ps” in the current working directory. One can import data from an ASCII file, where columns are separated by white space.

So that’s exactly what I did. 5.2 million data points in, one 151MB Postscript graphic out. Ghostscript can convert that to PDF, which can then be converted to all sorts of raster-based graphics formats. It’s not a perfect solution, but it is a working solution.

Back to the grind for me…

Back in 2006, we passed on a computer system to a friend of ours who had none at home. Since then, she has gotten interested in digital photography and wants to produce some designs for CafePress, as well as the usual email, browsing, and writing. This is complicated by the fact that she and her husband don’t really care to spend the bucks for broadband.

Yesterday, she was shopping and ran across a sale on a Compaq Presario F750US laptop computer, at $450. She ended up buying it after consulting with Diane and I. She is planning on using some of the free WiFi hotspots in her vicinity, such as the local branch of the public library, in order to use the Internet.

We talked a bit later, and I mentioned that in keeping with the budget she’d like (spend as little as possible), that there were a number of open source or free applications and utilities that she should look into. I’ll append what I wrote to her. Please add other suggestions or disagreements on my suggestions to the comments.

(more…)

Part of the Antievolution.org domain is an online web-based bulletin board system; let’s call that the AE Public BB. It has been operating since about 2002 with the current software, the Perl-based IkonBoard system. IkonBoard has pretty much become a legacy system without much in the way of support or user development, but the instance I have has had various modifications that provide useful features. Back when we were having issues with putting too much load on our old server, the IkonBoard installation was the application that showed the fewest problems, though even it was affected.

The point of having a BBS is to get a good community of commenters, and this the AE BB has done, in large part due to its being the spillover discussion area from the Panda’s Thumb weblog.

Early on, my idea was that threads on the BBS would bring together interesting and useful knowledge about topics, which could then be hosted as resources in themselves. I even made a start then of hand-saving the HTML of some threads to put in the main AE domain space. But at the time the BBS was using a “dbm” style database, and my only access to it lay through the IkonBoard interface. Since then, I’ve switched the database to MySQL and programmed a couple of accessory pages for mining information from it.

IkonBoard, like many other BBS systems, uses an arcane URL structure to specify threads on the BBS. It also insists on inserting a session identifier into its URLs generated dynamically. All of those things tend to make search engine spiders either give up entirely, or for the backends to give little weight to pages that do manage to get indexed. While I’ve added a hack that presents search-engine-friendly URLs to search engines, it still seems like pages don’t get indexed, or at least don’t reliably show up in searches.

So, given the early intentions and the more recently acquired means of doing something about it, I’m pleased to unveil the AntiEvolution.org Public Bulletin Board Archive, a page that presents the content of the AE Public BBS via links with meaningful names (they are taken directly from the topic titles). The procedure behind it will make it simple to update this archive periodically, and I’m intending that to be every month or two. There is a web-interface application in PHP that queries the MySQL database to find all topics that are cleared for public viewing. That application builds a page that has the HTML for a links page, and a set of commands for retrieving each topic as a single (sometimes very long) page from IkonBoard. Because IkonBoard generates each page, all the original links work, and the theme carries over. A user interacting with the pages will be shifted over to the IkonBoard BBS system and its dynamic, up-to-the-minute content.

So I think that this will serve the purpose of saving the sometimes useful, and more often entertaining, exchanges that take place on the BBS. There are definitely things there that should be remembered, and not be treated as evanescent and disposable interchange.

Remember a while back when I talked about HP’s Media Home Server, a way-high-price-point box that runs the Windows Home Server operating system? It turns out that there is something else that might make consumers think twice: it is possible (though it sounds fairly unlikely) to get into situations where you think that you are backing up your files to the server, but actually you are turning them into corrupted garbage. This brief CNet article breaks the bad news. It sounds like the system dealing with files can get confused if overloaded.

I hope that Microsoft is able to fix this quickly; potential data loss is a show-stopper in my estimation. I am trying to recall any instance of data loss in the various Unix-based servers I’ve run or used over the years, and pretty much any time I’ve lost data, it’s turned out to be hardware failure or my own fault. Here’s an article detailing exactly how to get up and running with a home server based on Linux about as painlessly as possible.

There’s a website that encourages people to toss in their blog URL and get a readability assessment. I did that, and got this:

I did snip off a second URL to the HTML they provided, something about “Cash Advance Loans”. Somehow, I doubt my elementary diction attracts the sort of people needing that service. But it does point out the good that can come from paying attention to HTML code for badges and other little bits that have become a part of blogging.

I heard about DesktopBSD from Skip Evans. This is a configuration of FreeBSD that provides support for automatic detection of hardware, especially that needed to get X11 working properly. It uses KDE as its window manager, and much of the work done integrates KDE with services of FreeBSD. This sounded interesting to me, since I’m used to snappy auto-detection by various Linux distros, but getting FreeBSD running X11 has always required a fair amount of geek work and worry whether one has gotten the actual video card and monitor values right. At least within memory, the graphics subsystem is the place in computers where getting something wrong could actually start a fire in a system.

I downloaded the AMD64 1.6RC3 DVD of DesktopBSD for a trial on my laptop system. This model, a Gateway MT6458, isn’t on the list of known good installs, but I figured it was worth trying. I booted into the Live mode, and, as I expected, the auto-detection worked for everything but the built-in wireless and the sound system. My Linux install, Xubuntu, doesn’t know about those, either.

I didn’t want to disturb my current setup, so I plugged in my external 2.5″ USB drive. It has a 14GB partition for FreeBSD, and the rest is formatted as a FAT32 data partition. I dropped out of the Live mode and started an installation. The menus brought me to selecting a drive, and I selected the external drive. Then there was selecting a partition, and I selected the old FreeBSD partition. I told it to install to that. There was the notice that changes would be permanent, and I was fine with that. A brief whir, and an error message popped up saying that the install was aborted, that “bsdlabel” had failed. Well, OK, maybe I need to let it try to work to a blank hard drive.

Except that on reboot, my external USB drive was no longer recognized. Both the FreeBSD and the FAT32 partitions had vanished. Ooops.

Well, there were things on the FAT32 partition that I was going to miss. I have looked into data recovery stuff in the past and found the commercial ventures offering such to be priced out of range of my wallet. I decided to search for “fat32 recovery “open source”", and one of the things that popped up was TestDisk. Open source, runs on just about anything, handles just about any format… sounded like just the thing. The Wiki for TestDisk handily provides examples. I downloaded the Windows binary, copied over the indicated directory to my Vista installation, and ran it. It found my main hard drive and the external USB drive with no problem. Asking it to analyze the external drive brought results in seconds: it found both the FreeBSD and the FAT32 partitions on the drive. All that looked fine, so I moved on and was prompted to write out the partitions. It did that, I rebooted the laptop, and am now getting stuff backed up so that maybe I won’t have to miss things if I do silly stuff like experimenting with storage media that have unique data on them.

Oh, DesktopBSD guys… a little more attention paid to not doing bad things to partitions on installation failure would be good. I’ll try again, but not with any partition at hand that I care about.

My Nikon D2Xs has an orientation sensor, so if I take a vertically-oriented photo, that information gets placed in the EXIF data in the JPEG file. Depending upon my viewing software, I may be presented the photo horizontally, or in the correct vertical orientation. Since I often am looking to do batch processing of all the files taken at a particular event, it would be nice if the EXIF data could be applied before making resized pictures, where I may need to manually rotate three versions of the original.

Enter JPEG-EXIF AutoRotate, a software package that brings together several utilities to do this job. When installed, this adds several options to the context menu under Win32 operating systems. One can ask for a file to be rotated, or selected files, or selected folders, possibly including subfolders. Best of all, the rotation is performed without incurring a loss due to JPEG de-compression/re-compression cycles. So now, at least with my most recent camera, I can reduce some of the workflow: download, autorotate, then run my resize script.

The only thing better would be if they included a way to invoke it from the command line. I haven’t dug through all of that site, though, so maybe they mention something of the sort there.

Something else that would be useful would be if someone could do the same job for Mac OS X and Unix systems.

Update: I should have noticed that a command prompt pops up when autorotation is selected from the context menu. I still apparently need to either copy the relevant executable files to somewhere in my path, or set the path to include their current installation directory, but I feel I’m close to having that working.

Update: The elements of doing this via the command line are there. Three utilities, “jhead”, “jpegtran”, and “mogrify” from ImageMagick are installed by the software I linked above. The context menu items launch batch files. The relevant command for autorotate of all JPEGs in a directory is then:

“C:\path to\jhead” -autorot -ft “dir with jpeg files”\*.jpg

So now I have this incorporated into my general resize batch process. It launches autorotate, then handles resizing, normalizing, and unsharp mask at three different final image sizes.

I was looking into information about plugin development, and happened across a reference to both the Sansa e280 MP3 player and something called RockBox. I have a Sansa e280, so I had to have a look into what the RockBox thing might be.

RockBox is an alternative, free, open source firmware for several different brands and models of MP3 players. The models that seem to be best supported are made by Archos, but the Sansa versions were listed as “Usable”, so that sounded like something to check out.

Working from the RockBox for Sansa user manual, I eventually got an installation going. I had tried initially to do that on the Mac laptop, but for some reason the Mac refused to recognize the Sansa more than once. I shifted over to the Windows laptop, and was able to get the installation completed and a run of “sansapatcher” to update the boot system on the MP3 player. I also selected a number of themes, including some featuring large type.

Well, it works, all right. I’m still getting used to its navigation scheme and its focus on playlists, but I figure that will come with some experience. Some things that I find better immediately are that the buttons and slide wheel are recognized instantly. The Sansa firmware apparently requires a slide or button press event to “wake up”, which can make things ambiguous. I’ve also installed a “talk” file, which means that RockBox “speaks” the current menu selection as I navigate menus. This means that I should not have to pay attention to small type on the little screen if I need to change something while doing other tasks.

I’ve also checked, and I can get right back to the Sansa firmware by holding down the “left” button while powering the unit on. Nice.

I’m using the DockPod Aqua Large theme currently, which gives me large type. I also set the foreground color to something very close to black. Full black is flagged as an invalid color for some reason.

There are a number of features that I hope to test soon. Rockbox provides support for displaying JPEG images and MPEG video. I may set up essentially a small portfolio and snapshot set. The video option might be good for air travel. It also comes with a number of games, including ports of classic arcade games like Asteroids and Qix, as well as chess and solitaire. It even provides several applications like “Clock” and “Calculator”.

If you like getting more out of your MP3 player, I think RockBox may be worth trying.

Update: There are two problems that I’ve noticed so far. First, the USB connection doesn’t happen under Rockbox; I have to reboot into the Sansa firmware to be able to transfer files. This is a known issue, and folks at the RockBox site are working on getting USB functionality going. Second, and far less important, is that my installation doesn’t seem to handle the Doom port. I’d like to have that working just for the nerd factor.

Something that I had some trouble with was getting Avida running on some of the operating systems I’m more familiar with, chief among them FreeBSD. This evening, I managed to put the pieces together for running Avida on FreeBSD 6.2. Of course, I’m checking out the development branch from the Devolab svn repository, so I’m not talking about the FreeBSD port of Avida.

Avida uses the Cmake project build tool, so the first thing needed is is to turn to the ports system and get current versions of Cmake and other dependencies installed. If one installed the “development” distribution in the FreeBSD install, you’ll already have most things. You will definitely need the GNU GCC compiler with G++.

Once you have a checkout of the Avida source, one needs to run “ccmake .” in the top-level directory of the tree. I’m just setting up targets of “avida” and “avida-viewer”, with “Debug” mode set. Once that works, one can launch the build process with “make”. That will churn for a while and crap out about half-way through. The issue? “log2″ is not defined in one of the source files. You’ll need to edit “source/main/cTaskLib.cc” and look for “log2″. You’ll find a #define for it in a conditional for a Windows installation. Copy that line and paste it in outside the conditional. Re-start the make. Everything will be fine thereafter. (tcmalloc throws a lot of warnings. It does that everywhere, though, so just ignore it.)

I’ll see about getting a conditional around that for FreeBSD in the main distribution.

Product News: HP MediaSmart Server Makes Big Debut, by Rachel Cericola - Electronic House Product News

CNet review of HP MediaSmart Server

HP has announced a computer system to be powered by the MicroSoft Home Server software. From the description, this is a networked PC with a big hard disk, so audio, video, graphics, and ordinary files can be stored on a central server in a local area network. There is some hint that this may be tied directly into features of other MicroSoft Windows operating systems.

The hardware itself is modest in its specifications: 1.8 GHz CPU and 512MB of RAM. A nice feature is four hard disk bays with sleds for easy addition and changes of hard disks in the system. As a server, it comes with a minimal set of ports: 3 USB, 1 e-SATA, and 1 Ethernet port.

What MicroSoft and HP offer with this system is the convenience of a pre-packaged, holds-your-hands solution for computer backup, media serving, and accessibility from the Web anywhere. They are charging $750 for a 1-TB model, or a 500 GB model for $600. Given the claimed ease of adding disk storage to the system, you’d be imprudent to pay the increment to get 1-TB; go to Newegg.com and buy a 500GB drive for a little over a $100 and add it to the system.

Back in 1998, I set up a home server based on the FreeBSD operating system. It provided central storage for files, interfaced with our Windows machines by Samba shares, ran servers to the internet, as well as serving as a router with NAT and firewall. This was before Linksys and others arrived on the scene with handy little boxes to do the basic routing for LANs through a internet connection. I used an old box with a 166 MHz CPU and, IIRC, about 64MB of RAM, and that worked fine. What was not provided was the dead-simple hook-up interface on the new stuff, which will make the new system useful for people whose lives haven’t included over a decade of professional system administration. I had to do some serious research to put all the pieces together on that first system.

We still use a FreeBSD server for our home LAN. It works on very modest hardware, and we can simply upgrade the box with whatever desktop machine becomes the next excess piece of hardware: shutdown, move the hard disk to the “new” system, and start it up again. So anyone who is modestly gifted with computer administration savvy can do much the same thing now with almost-free hardware, maybe adding the big hard disks for the job, and avoid the cost of the dedicated system being sold.

For those without the technical know-how or confidence, though, the MediaSmart server or equivalent from other sources may be something to consider. It costs as much as a low-priced laptop, though, which I think is going to limit the market for this sort of information appliance. When given the choice of buying infrastructure (server) or buying productivity or pleasure (new desktop or laptop), I’m thinking that many cash-limited families will go with productivity or pleasure.

Something I haven’t discussed so far is the bit about being able to access your data from the outside Internet via the Home Server. First off, this likely violates most user agreements that Internet Service Providers force upon users. This could be a hopeful sign, an indication that MicroSoft is taking a step toward empowering users, at least in the limited respect of backing their ability to use their own computers to serve their own data. Second, though, comes the doubts about security. Does MicroSoft Home Server come battened down such that you are the only likely person to be given access to your private data on your own machine? I haven’t seen this addressed in any review yet. Until such time as this is thoroughly vetted, I’d recommend any purchaser of these servers turns off outside Internet access. Assuming that the software allows them to do so… if not, then do it via settings on your router. If you are storing your financial data on this device, you definitely don’t want to be, essentially, inviting in identity thieves to help themselves to it.

Update: Thinking a bit more about this, I’d like to strengthen what I said above. I don’t think that this particular offering is going to do much at the price point stated. Consumers looking for a file server can buy Network Attached Storage devices at a better price point, or simple hard disk backup for about the cost of media plus $50 or so. I’m thinking that the breakover point between ho-hum sales and something that a whole lot of people will be willing to plunk down their money for will be somewhere between $150 and $200 over media cost, so for a 500GB Home Server system, around $260 to $310 total MSRP. The initial announced price is about double that. As it is, the system is competing with cheap laptops, mid-range desktops, prosumer digicams, video cameras, or flat-screen HDTV in the consumer’s budget. I think it is going to lose.

I happened to post an essay today concerning the unacknowledged errors in an essay by Robert Marks and William Dembski. In following up to a reader comment, I was skimming through another essay from the same website containing the one my post was about. Within its pages, I discovered that Dembski and Marks were treating readers to a description of a program Richard Dawkins discussed in his book, “The Blind Watchmaker”. Since the program seeks a target of a string, “METHINKS IT IS LIKE A WEASEL”, it is usually referred to as the “weasel” program. The program is simple enough to describe accurately, but Dembski and Marks botch the job thoroughly.

The accurate way to describe Dawkins’s “weasel” program is this way:

1. Use a set of characters that includes the upper case alphabet and a space.

2. Initialize a population of n 28-character strings with random assignments of characters from our character set.

3. Identify the string or strings closest to the target string in the population.

4. If a string matches the target, terminate.

5. Base a new generation population of size n upon copies of the closest matching string or strings, where each position has a chance of randomly mutating, based upon a set mutation rate.

6. Go to step 3.

Dembski and Marks claim that Dawkins’s “weasel” program smuggles in information of the target by use of partitioned search, by which they mean that every time a particular letter matches in a string, that correct assignment of letter and position are retained in all future generations. This is equivalent to the bad old password login scheme on DEC machines, where one could figure out which characters in a proffered password matched and continue login attempts until all of them matched. However, that is not how Dawkins’s “weasel” program worked, so every statement Dembski and Marks make that is based upon that false premise is utterly meaningless. Just as the Marks and Dembski “ev” critique failed upon easily checkable stuff, one can run Dawkins’s “weasel” and get results similar to those reported by Dawkins; however, running Dembski and Marks’s “weasel” with the partitioned search will return results that are far different from what Dawkins reported.

The “evolutionaryinformatics.org” website lists the paper in question as being “in review”. If it passes under the noses of sleepy reviewers somewhere, I look forward to having a corrective letter published in the same venue.

What makes this all particularly amusing is that Dembski knows that his description of Dawkins’s “weasel” program is erroneous. He knows about this error because I emailed the following text to him:

Information request to William Dembski:

[Quote]
He starts with a target sequence taken from Shakespeares Hamlet, namely, METHINKS IT IS LIKE A WEASEL. If we tried to attain this sequence by pure chance (for example, by randomly shaking out scrabble pieces), the probability of getting it on the first try would be around 1 in 10^40, and correspondingly it would take on average about 10^40 tries to stand a better than even chance of getting it.12 Thus, if we depended on pure chance to attain this target sequence, we would in all likelihood be unsuccessful. As a problem for pure chance, attaining Dawkinss target sequence is an exercise in generating specified complexity, and it becomes clear that pure chance simply is not up to the task.

But consider next Dawkins’ reframing of the problem. In place of pure chance, he considers the following evolutionary algorithm: (1) Start with a randomly selected sequence of 28 capital Roman letters and spaces (thats the length of METHINKS IT IS LIKE A WEASEL); (2) randomly alter all the letters and spaces in the current sequence that do not agree with the target sequence; (3) whenever an alteration happens to match a corresponding letter in the target sequence, leave it and randomly alter only those remaining letters that still differ from the target sequence. In very short order this algorithm converges to Dawkinss target sequence. In The Blind Watchmaker, Dawkins recounts a computer simulation of this algorithm that converges in 43 steps.13 In place of 10^40 tries on average for pure chance to generate the target sequence, it now takes on average only 40 tries to generate it via an evolutionary algorithm.

[End Quote - WA Dembski, "Can Evolutionary Algorithms Generate Specified Complexity", "Nature of Nature" conference, Baylor University]

There are several issues that this text brings up. Of the three steps listed as comprising Dawkins’ algorithm, only step (1) has anything like it in the pages of “The Blind Watchmaker”. Steps (2) and (3) appear to be inventions rather than descriptions. What is the basis for claiming that steps (2) and (3) represent Dawkins’ “weasel” algorithm?

Further on, the issue of “tries” it takes to find a solution is raised. For “pure chance”, a figure of ~10^40 “tries” is given, which would correspond to individual candidate solutions tested. For “weasel”, though, only ~40 “tries” are given, but in this case the number 40 derives from the number of generations taken by the “weasel” algorithm rather than the number of candidate solutions examined. It seems to me that for the purpose of comparison, a “try” ought to mean the same thing for both approaches. I would like to see a restatement of the section concerning “tries” that takes this into account.

Wesley

I even corresponded with Dawkins to make sure that there were no editions or versions of “The Blind Watchmaker” that incorporated anything arguably like Dembski’s inventions.

One may be wondering about the title I chose for this little essay. Well, the traditional anniversary gift lists says that copper and wool are the materials for a seventh-year anniversary. And my email to Dembski was sent:

Date: Mon, 9 Oct 2000 11:38:11 -0500 (CDT)
From: “Wesley R. Elsberry”
Message-Id: <200010091638.e99GcBg25785@inia.cls.org>
To: dembski@discovery.org
Subject: Information request re: Dawkins’ “weasel” algorithm
Cc: evolution@calvin.edu, welsberr

Back over the summer, William Dembski was talking up “Baylor’s Evolutionary Informatics Laboratory”, and one of the features there was a PDF of an essay critiquing the “ev” evolutionary computation program by Tom Schneider. Titled “Unacknowledged Information Costs in Evolutionary Computing”, the essay by Robert J. Marks and William A. Dembski made some pretty stunning claims about the “ev” program. Among them, it claimed that blind search was a more effective strategy than evolutionary computation for the problem at hand, and that the search structure in place was responsible for most of the information resulting from the program. The essay was pitched as being “in review”, publication unspecified. Dembski also made much of the fact that Tom Schneider had not, at some point, posted a response to the essay.

There are some things that Marks and Dembski did right, and others that were botched. Where they got it right was in posting the scripts that they used to come up with data for their conclusions, and in removing the paper from the “evolutionaryinformatics.org” site on notification of the errors. The posting of scripts allowed others to figure out where they got it wrong. What is surprising is just how trivial the error was, and how poor the scrutiny must have been to let things get to this point.

Now what remains to be seen is whether in any future iteration of their paper they bother to do the scholarly thing and acknowledge both the errors and those who brought the errors to their attention. Dembski at least has an exceedingly poor track record on this score, writing that critics can be used to improve materials released online. While Dembski has occasionally taken a clue from a critic, it is rather rarer that one sees Dembski acknowledge his debt to a critic.

In the current case, Marks and Dembski owe a debt to Tom Schneider, “After the Bar Closes” regular “2ndclass”, and “Good Math, Bad Math” commenter David vun Kannon. Schneider worked from properties of the “ev” simulation itself to demonstrate that the numbers in the Marks and Dembski critique cannot possibly be correct. “2ndclass” made a project out of examining the Matlab script provided with the Marks and Dembski paper to find the source of the bogus data used to form the conclusions of Marks and Dembski. vun Kannon suggested an easy way to use the Java version of “ev” to quickly check the claims by Marks and Dembski.

(more…)