These are my thoughts on the .LRN learning management system as of version 1.0.1.
The system is not straightforward to install. It inherits all the complexities of the underlying OpenACS web application server. The installation instructions themselves are fairly clear and accurate, but just a bit long and complicated. This is probably due to the long list of packages that .LRN depends on.
My first impressions of the system out of the box is that it's not very pretty. The standard theme is just a little bare and uninteresting. Compare with the PostNuke standard themes which, besides being numerous, look quite rich.
It only takes a few minutes to browse around the default empty site and get a feel for the user interface and the capabilities. What eluded me at first, though, was: where was the administration configuration? And by administration, I don't mean the administration of the courses, the users, and files. I mean the grand poobah of admins: the one that installs the .LRN packages, that changes the name of the web site, the theme. It turns out that some--not all--of those features can be found if you point your browser at http://yoursite.com/index.adp, which brings up a page different from http://yoursite.com/. This is not a documented feature for some reason. From this page, you can find API documentation, install packages, etc.
For my purposes of creating a community of self-administering self-learners, .LRN is too rigid and lacking in features. Once .LRN 2.0 comes out in the coming weeks, I might take another look at it. In the meantime, Yahoo! Groups will just have to do.
Wired News: Pop-Up Scam Beats AOL Filter tells us which ports to block to avoid annoying Messenger spam pop-ups on Windows:
AOL is not the only Internet service provider currently blocking all port 135 traffic. Many ISPs began filtering the port last month to mitigate the spread of the MSBlaster computer worm, Baldwin said. While AOL also could block UDP ports 1025-1029 to fully eliminate Messenger service spams, the big ISP likely is worried about the potential "collateral damage" such a move might cause to users with legitimate programs that require those port addresses, he said.
As per .LRN 1.0.1 installation instructions, ran:
./configure --with-aolserver-src=/tmp/src/aolserver/aolserver --with-tcl=/usr/lib/tcl8.3
But when running make, the build system couldn't find the tcl header files.
I found the easiest solution was to symlink all the files I could find in /usr/include/tcl8.3 into this source package's include directory. [NOTE: I later found that these header files are in the aolserver/aolserver/include directory]
Unfortunately, another step in the installation requires the building of supplemental modules which must be constructed from a PostgreSQL source tree that includes the contrib directory. These sources, I don't have available since I have a Debian package installed. But even if I had the corresponding sources, I don't even know if PostgreSQL 7.2.1 would even work for these installation instructions, which presuppose 7.2.4.
At this point, I'm going to postpone an attempt at installing this feature. When .LRN 2.0 comes out, it's possible that it will support PostgreSQL 7.3 in which case we can install the Debian package (of testing status) which might already include this contributed module pre-compiled.
Just followed .LRN 1.0.1 installation directions.
I installed in /usr/local/aolserver as directed.
Modifications necessary:
- Contrary to my normal in-house installation procedure, I did not run my homegrown lnsrc script. The symlinks created by lnsrc make the installation create incorrect destination files.
- I had to modify conf-make to contain "make" instead of "gmake".
Compilation error encountered:
After compilation, I checked the build logs in the log directory and found that nspostgres did not compile right:
nspostgres.h:48: libpq-fe.h: No such file or directory
make: *** [nspostgres.o] Error 1
It turns out that I need the posgresql-dev Debian package installed, as per this AOLserver mailing list thread. I also had to unset POSGRES environment variable, unlike what the instructions say. And I had to fix the nspostgres Makefile, which has PGLIB and PGINC switched in one of their definitions. This will force the Makefile to figure out where the include files are.
I tested the AOLserver as per instructions (with one difference being to use the 'www-data' group instead of the 'web' group).
This is my attempt at installing database PostgreSQL 7.2.1 on Linux Debian 3, as required by the Learning Management System .LRN 1.0.1 which is built on the web application server OpenACS 4.6.3. Apparently, on my machine someone has already installed the PostGreSQL 7.2.1-2woody2 package. But the post-installation script was not run; in other words, no functioning database is up and running.
Upgrade
There was a 7.1 database installed on the machine, which had not yet been updated and which no one was claiming. Instead of letting the postgresql post-installation script update this probably useless database, I just moved the /var/lib/postgresql/data directory out of the way.
Default database
Then I unleashed /var/lib/dpkg/info/postgresql.postinst configure, which created the default template1 database.
From now on, the database should be automatically restarted upon booting via /etc/init.d/postgresql start
.LRN/OpenACS installation instructions
I followed .LRN installation instructions with respect to postgresql.
Added plpgsql support to template1:
createlang plpgsql template1
createlang -l template1
Actually, it was already done.
Tested PostgreSQL by creating a temporary database.
Tuned PostgreSQL.
This is my attempt at installing the Learning Management System .LRN 1.0.1 on Linux Debian 3.
Prerequisiste Software
Software installed:
- .LRN 1.0.1 which includes OpenACS 4.6.3
- tcl8.3-dev Debian package (already installed on my machine)
- libxml2 and libxml2-devel Debian packages (already installed)
- AOLserver 3.3oacs1 (instead of the AOL server 3.4.2-1 package available because the offered version should have all the patches needed by OpenACS.)
- OpenFTS TCL 0.3.2
- Analog 5.23-0woody1 Debian package (already installed)
- PostgreSQL 7.1.2-2woody2 Debian packages, including postgresql-contrib and postgresql-dev (partially installed)
To be installed later:
- Analog 5.32 Debian testing package
- Daemontools 0.76
.LRN installation
Created the lrn user.
Created a /usr/local/openacs/web directory for OpenACS services.
Created the lrn user with that directory as its home.
Did the CVS checkin and -out.
Created the appropriate directories.
Created postgresql user and database.
Configured AOLserver.
Configured the AOLserver service for .LRN by using the OpenACS Installer through the web, allowing the database to be initialized.
Backed up the database.
Updated CVS
Installed cron job to back up the database daily.
Set up analog.
Linked images with this command:
ln -s /usr/share/doc/analog/images /usr/local/openacs/web/lrn/www/log
Running through proxy
Because my machine already had an Apache server listening on port 80 and serving other users on the machine, I had to use a proxy to redirect to the OpenACS web server.
Here are my changes to the apache configuration:
<VirtualHost 25.219.351.34>
ServerName openyourmind.com
ServerAlias *.openyourmind.com
ServerPath /openyourmind
ProxyPass / http://127.0.0.1:8000/
ProxyPassReverse / http://127.0.0.1:8000/
</VirtualHost>
And I had to modify config.tcl in /usr/local/openacs/web/lrn/etc/config.tcl to set address 127.0.0.1. Don't know why I had to do that.
Note that 'localhost' didn't work in place of 127.0.0.1.
Of course, I also had to set up DNS to accept this virtual host.
Some people view the phenomenon of Fakesters as a risk to the Friendster vision of an effective friend-of-a-friend dating network. Others abhor some users' tendency to add random people to their friendster lists to augment their connectivity. I believe that all these problems indicate social urges that might be satisfied cleanly, were not for Friendster founders' tunnel vision.
Fakesters
People are not only being creative with their invention of fakester profiles. They're also screaming out that Friendster is lacking in more complex ways to form social networks . The idea that friendship is the only type of social relationship that exists in this world is simply dumb. The solution is for Friendster to allow for the creation of groups or clubs that anyone can then associate with.
Friendster hubs
If your friendster list is parse, it might make sense for you to find a well-connected individual or two and request them to become their friendster. This could greatly enhance the size of your Personal Network. Why would anyone do that? Well, simply to increase the results of one's user searches, thus allowing for more possible encounters. If this is a dating site and dating is competitive, you have to expect users to do this.
To discourage this kind of behavior, which arguably pollutes the meaning of the friendster relationships, the solution is just to allow for everyone to determine the way their Personal Network is calculated. In other words, an option should be given for users to specify the maximum number of hops allowed in searches. Why arbitrarily limit who users can meet?
A similar option can be done for the other end of a search, to regulate privacy. Everyone should be able to specify the maximum number of hops allowed before her profile is revealed to the searcher.
w.bloggar is a nice piece of software. Just got a few complaints:
- The account login screen doesn't have a Windows task bar entry or a system tray icon. You can't even access it using the ALT-TAB key sequence. So if the login screen is hiding behind your other windows, you're going to have to minimize them one at a time in order to find it.
- There needs to be some sort of account refresh. If I add a category in Movable Type or update its name, I'd like to be able to refresh w.bloggar's listing so that I can post appropriately. Clicking on Select Account and logging in again is obviously not the best answer.
- If you change your toolbar configuration, you lose your layout as soon as you post.
After following version 2.64 upgrade instructions and pointing my browser to mt-upgrade25.cgi, I get:
Running 'alter table mt_author add column author_preferred_language varchar(50)'
An error occurred while upgrading the schema:
alter command denied to user: 'mt@localhost' for table 'mt_author' at /home/www/mt/mt-upgrade25.cgi line 61.
I give up.
DVD+R vs. DVD-R
When deciding whether to buy a DVD+R or DVD-R drive, it should be noted that DVD+R is becoming the format of choice these days. It is technically superior to DVD-R. It also ships with DELL PCs, a significant factor in predicting de facto standard winners. While you may buy a dual-format DVD burner, you can rest assured that buying a DVD+R burner will be fine.
As for media prices, there does not seem to be a difference in price between the two types of blank DVDs nowadays.
PAL vs. NTSC
My first DVD was in fact created in PAL format. One thing to realize is that most DVD players will happily and automatically convert between the 2 formats NTSC and PAL. So even in the US you can easily test your PAL creation in your home DVD player without worrying about funny things being shown by your TV.
Sonic Foundry's DVD Architect 1.0 is a great choice for those brave souls wanting to dabble in DVD authoring at home. The user interface is intuitive and makes DVD authoring seem like a breeze. But due to the newness of this product and DVD technology in general, there are bugs and quirks in the product that you will have to sort out before you reach your DVD goals. Here's a sample of my problems. I'm sure there are many more waiting to be discovered.
Slide show
A nice feature of DVD Architect (DVD-A) is the ability to easily create a slide show out of a number of photos. It's important to realize that, when creating a slide show, a running video is essentially created. This video consists of the sequence of photos with a predetermined pause in between. So the longer the pauses between the photos, the more space the slide show will take. I found that for my taste 3 seconds was the right balance between speed and size; your mileage may vary.
Also, a bug in earlier builds of DVD-A prevented rotated photos from staying rotated when the project was saved. An update to the patch 1.0c solved that problem.
Estimated Size in Optimize DVD
The first thing you will notice when running DVD-A's Optimize DVD function is that the estimated size is compared against 3.95 GB media instead of the more common 4.700 GB media. The solution here is to go to the DVD-A Options.
Then you might be wondering what GB actually means. Does it mean gigabytes as we traditionally know them or billions of bytes? And how close can we safely optimize to 100% of the media size?
Trying to figure those things out was impossible because the numbers just did not make sense any way I twisted it. Until I installed the 1.0c update. This update definitely changed the way the estimated size was calculated. But that still doesn't answer our questions. Well, after burning one CD, it became somewhat clear that MB refers to millions of bytes and GB refers to billions of bytes.
In other words, it is safe for you to optimize all the way up to, and including, 100% of the media size, as stated by DVD-A.
(NOTE: the numbers still don't make complete sense as you would expect the estimated size to say 4700MB when you optimize to 100%, but instead it says 4812MB. This would seem to indicate that MBs are still referring to megabytes. But I recommend you just ignore this quirk and just rely on the 100% measure.)
Bitrate in Optimize DVD
A newbie to the DVD scene might wonder what bitrate to choose when optimizing a DVD in DVD-A. I tend to be fairly lenient when it comes to video quality and I found that 3 Mbps, instead of the default 8 Mbps, worked fine for me. This allowed me to fit 2 hours of video and 5 photo slides, including hundreds of photos spaced 3 seconds apart, onto a 4.7GB DVD+R.
Burning DVD+RW
When trying to burn a blank DVD+RW, DVD-A will burn the lead-in and then simply stop without burning the actual content. It will then claim success. There's no bug fix at this point. But the workaround, as suggested in the Sonic Foundry forums, is to burn some sort of sample DVD, using for example Sonic MyDVD. Then you can let DVD-A burn over your DVD+RW.
In the end, burning a DVD+RW still failed for me, as certain chapters in my DVD could not be found by any of my players. I simply turned to a demo version of Ahead's Nero Burning ROM to burn the DVD files prepared by DVD-A, which did the trick. I don' t know if this problem applies to DVD+R because I'm not going to burn a write-once DVD unless either I can successfully burn a DVD+RW first or the price of DVD+R media drops significantly.
