crevier's blog

I'm working on implementing a blog solution for students to be marketed by our admissions office. Gotta have something that allows students to write in their own blog and have a moderator approve/publish the posts. Having authentication via LDAP is also a big plus. Also prefer Linux/Apache/MySQL/PHP/Perl. I'm currently playing with b2evolution, though I'm interested to hear about other possibilities.

Since there seems to be some pretty decent free/open source solutions out there, that's where I'm focusing my efforts right now.

So now I'm developing a financial aid calculator for our admissions office. It will allow a potential student to enter their GPA from high school along with either their ACT or SAT score. Then our web site will provide a financial aid dollar amount along with some verbiage from our admissions staff. So far it's been interesting.

Spending a little time looking at other colleges to see how they've done it.

Mozilla released its version 1.5 of Firefox last week. And like thousands of techies everywhere, my first thought was, "will it work with all of my extensions?"

I manage the technical aspects of our web site, and I've been wanting to setup a good, reliable link checker for some time. So I just implemented Linklint last week. It's a freebie written in Perl, and it works great. It's fast too! It's very configurable, so it took a bit of effort to set it up. But now it's in a cron job, creating daily reports by 8:00am.

Actually, one of my favorite link checkers has always been Xenu. That one is super fast and very accurate. But it only runs manually as a Windows program. There's no way to schedule it, and it won't run on my Linux server. But I still use it if I need to run a quick and dirty report on a particular web page.

Holy cow, what a month! I've come to learn that August is pretty busy around here, preparing the campus and our network for the arrival of students.

I was just getting used to posting occasional blog entries, and then August hit. It was pretty typical around here, but on top of all the usual exciting activities, we also launched a new content management system for our web site. We implemented Cascade Server (formerly ContentXML) and all went pretty well. I've got several campus departments now maintaining their own content using the new system. We're going to be adding our academic programs web content to the system next, and 2006 will be the big year to get the entire campus using it.

Through the use of the CMS, I'm now able to give certain people access to update certain parts of criticle pages, like our home page. I also created a MySQL database for our press releases as well as our A-Z index. Now, a non-techie person can maintain those databases, and the content automatically appears on the web site via a couple of CGI programs that I wrote.

This topic came up on a discussion list today, so I figured I'd post some info about it. If you have images on your web site that you don't want displayed on other web sites, and you're running Apache, you can easily prevent it. I don't know about a similar solution in other web servers.

This solution does not refer to the situation where someone else actually downloads your image (which is practically impossible to prevent), puts it on their web site, then displays it on a page. It only addresses the situation where someone else refers to the full URL of your image in the the HREF element of the A tag on their page. This of course creates unwanted bandwidth for your site.

Apache allows you to prevent this using the RewriteEngine. Instructions are available in their URL Rewriting Guide. Just find the words "inline-images" on the page.

You'll notice that the solution depends on the browser providing a proper value for the HTTP_REFERER environment variable. Some may say that this is not reliable, and that's indeed true. But while it's not 100% reliable, it's certainly reliable enough for this situation. Think about it. In order for this solution to not work, the browser (user agent) would need to send the wrong data for the referring site. And since the browser is in the control of the person visiting the web site, it's completely out of the control of the web site that wants to display your image. So the solution works very well.

I'm interested to know if there are similar solutions for other web servers.

I just made some source code available on our web site last week that I thought others may be interested in. We use NetReg to require all computers on campus to register before gaining access to the campus network (and the internet). Last summer (2004), we were trying to figure out how we can enforce some minimum requirements on our student computers before allowing them to register. For example, we make McAfee Virusscan available for all students, but we didn't have a way to require them to install it.

So, I wrote a small program (using AutoIt) called SNC Validator to accomplish this. When a student first fires up their computer, they only have access to one web page, our page that allows them to register their computer. This web page now allows them to download SNC Validator (a 124KB exe file) and run it. The program reads some registry settings (for Windows version, SP level, installed apps like McAfee, etc), builds a long URL, then lauches that URL in the default browser. By launching the URL, it is sending info about the computer to our NetReg server. Then, the Perl CGI script on the NetReg server analyzes all of the data and decides if the computer can be registered or not.

For example, if the CGI script sees that the user is running Windows XP, it can also make sure the user has SP2 installed. Or it can require that a certain version of McAfee Virusscan is installed based on the OS. If a specific requirement is not met, the CGI script prints a message telling the user what they need to do (i.e. install McAfee). Once the user does that, they need to run SNC Validator again, which reads the registry again and the process keeps going until all the requirements are met. Note that the SNC Validator program makes no decisions. It simply gathers and reports information. The Perl CGI script on the NetReg server is what makes the decision.

The nice thing about this setup is that the register/no register decision can be made based on any registry settings, the existence of any file, etc on the student's computer. But it does require some programming with AutoIt and Perl. If you haven't used AutoIt yet, you gotta check it out; it's quite handy for many other uses too (and it's free).

If you're using NetReg and want to give this a shot, the source code is available on our SNC Validator web page. This is only the source for the Windows EXE file. It's my hope that based on this setup, you'd be able to edit the CGI script yourself to incorporate whatever requirements you have for your campus.

We've been using this for a year now, and it has improved the overall level of security on our student computers greatly.

Five weeks ago, I returned from a one week vacation to find a dead hard drive in my computer. It's a standard DiamondMax Plus 8 40GB drive that came with my HP/Compaq WinXP Pro system. When I booted up, it just did some clicking and said that the drive had 0 MB on it. We have since sent the drive to DriveSavers for recovery, but they were not able to recover any data.

I think they normally charge a couple hundred dollars for an attempt fee, but for us, they charged nothing because we're an educational institution. The recovery fee was estimated at $1700. So, since they would not get any payment unless they recovered data, and since they sent it back with no results, I'm assuming that my chances are slim at finding someone who can recover the data. We've now sent two such drives to DriveSavers with no luck. I'm wondering if anyone else has had any luck in this area, with DriveSavers or anyone else.

(I've spent the past 5 weeks building a new life on my new hard drive. I've been working here for a year and a half and no, I didn't have any backup of any kind. It's like being a new employee again. No e-mail. No bookmarks. I've already heard it all from my loving and caring colleagues here at SNC, so don't start.)

I'm looking for a good blog solution for use by students, faculty and staff. It's got to run on Linux with Apache. Other positives would include MySQL, Perl, PHP.

An important component we'll need is LDAP authentication. Or, I already have my instance of Apache using LDAP for its standard HTTP authentication, so if the blog solution supports HTTP authentication, but not LDAP specifically, that would be fine.

I've been looking at WordPress. I'm wondering if there are others I should be checking out.

ContentXML makes good use of XSL, so I started playing around with XSL and XSLT today for the first time. This is some pretty slick stuff. If you haven't seen the tutorials at W3 Schools, you should check them out. They take you through the very basics of creating an XML file along with an associated XSL style sheet. Where have I been for the past 6 years?