So what is a tagged PDF? Well, it’s a PDF that contains structural information about the content, e.g reading order, the presence of tables etc. This allows screen-readers to read the PDF document – it makes the PDF accessible. In order to get the most out of the conversion process, as much structural information as possible needs to be present in the original document. I came across these recently when doing some work for a local authority that is using Alfresco.

So how do you configure Alfresco to produce tagged PDF? Open up the file ‘openoffice-document-formats.xml’, which is located in <tomcat_home>/webapps/alfresco/WEB-INF/classes/alfresco/mimetype/, locate the Portable Document Format document format section (it should be at the top of the file) and modify it so it looks like this:

<document-format><name>Portable Document Format</name>
<mime-type>application/pdf</mime-type>
<file-extension>pdf</file-extension>
<export-filters>
 <entry><family>Presentation</family><string>impress_pdf_Export</string></entry>
 <entry><family>Spreadsheet</family><string>calc_pdf_Export</string></entry>
 <entry><family>Text</family><string>writer_pdf_Export</string></entry>
</export-filters>
<export-options>
  <entry><string>EnableTextAccessForAccessibilityTools</string><boolean>true</boolean></entry>
  <entry><string>UseTaggedPDF</string><boolean>true</boolean></entry>
</export-options>
</document-format>

Restart Alfresco. That’s it! The next time you convert a document to PDF it should be tagged. You can test that the conversion worked (on the Mac) by using Adobe Reader 8.0. Open up the PDF file. Go to Document -> Security -> Show Security Properties. Click on the ‘Description’ tag. The ‘Tagged PDF’ entry should be set to ‘Yes’ if the conversion worked correctly. You can also check the document for accessibility by clicking on Document -> Accessibility Quick Check.

You can download the modified configuration file here.

By default Alfresco will convert PDF, Word documents etc to plain text but if you want to go the other around, e.g. transform plain text to PDF, you need to start Open Office from the command line – a version of Open Office comes bundled with Alfresco. It’s easy really: just run the start_oo.sh script, which you will find in the Alfresco home directory. Once you have done this you should be able to transform documents to PDF.

You can check that Open Office is running by executing the command lsof -i | grep 8100. If you start Open Office manually, note that Alfresco expects Open Office to be running on port 8100.

There’s a whole page on starting Open Office from the command line on the Alfresco wiki. Not sure why the page needs to be so long. I just ran the start_oo.sh script and Open Office started up without any problems. Maybe I got lucky!

Anyway, it turns out that TRAMP is not included with the version of Emacs that I had installed (sigh). To install it type “sudo apt-get install tramp” on the command line and restart Emacs. Viola!

To set the default connection mode to ssh, edit your .emacs file and add the following line:

(setq tramp-default-method “ssh”)

I guess it could be because I don’t take enough breaks away from the keyboard. However, yesterday I found an application for Mac OS X that reminds you every so often to take a break from typing. It’s called AntiRSI. It runs in the background and every so often up pops a window telling you to take a break. It gives you a great excuse to go and get a coffee!

David has more information on his site here. His article also includes a link to an application for Windows that works just like AntiRSI called Workrave.

tcpflow -c -i <interface> host maps.google.com

The -c flag means print to the console.

It’s a useful tool and it means that I don’t have to figure out how to use tcpdump! You can download tcpflow here.

1. Create the log file

   
   > su
   > cd /var/log
   > touch mysqld.log
   > chown mysql mysqld.log
   

2. Edit /etc/my.cnf

   
   [mysqld]
   log = /var/log/mysqld.log
   
   

3. Restart the database

If you installed the MySQL Startup Item then you can restart the server simply by running:

sudo /Library/StartupItems/MySQLCOM/MySQLCOM restart

Job done!

On a related theme, I often see broken images when viewing RSS feeds in news readers — once upon a time I had the very same problem in my feed. The reason: relative URLs.

For example, let’s take the image of myself. If you looked at the HTML for this particular image you would see something like this: <img src=”/images/me.jpg”/>. Works fine in a browser but try it in a news reader and … broken image. To view the image in a news reader the link should be: <img src=”http://simonbuckle.com/images/me.jpg”/>; notice the inclusion of the (full) domain name. Perhaps that’s another idea for a WordPress plugin: for each blog post add the domain name to any relative URLs to avoid things like broken image links.

In summary: check how your feed looks in a news reader! Your readers will thank you or, at the very least, you won’t look like a dumbass.

Here’s a solution I came up with: use the IP address of the originating server the mail came from and do a reverse DNS lookup to get the domain name. Compare the domain name from the lookup with the domain the email claims to have been sent from and if there is no corresponding A record discard the email. How reliable is this method? I have come across several references on the Internet that say that looking up the PTR record for a domain is not reliable because the PTR record(s) are often not well maintained or are even absent.

Another way using a similar approach but not relying on PTR entries, would be to look up the IP address of the domain the email claims to have come from and compare that with the IP address of the originating server. Have I missed something? Are there better ways of checking for spoofed emails?

Your comments are appreciated.

Add the following Javascript in the URL field at the top of the browser and then bookmark it:

javascript:self.moveTo(0,0); self.resizeTo(screen.availWidth, screen.availHeight);

Add the bookmark to the bookmark bar, click on it and voila …. Safari occupies the whole of the screen. Opera has a better full screen mode, it even covers the menu bar at the top of the screen, but until Apple add such a mode to Safari this will have to do.

Here is snapshot of my toolbar:

As I mentioned previously there is a plugin available called Saft that adds, among other things, a full screen mode to Safari. According to the authors of Saft it was used for check-in at Macworld San Francisco in 2005. Kind of makes you wonder why Apple doesn’t just clone the functionality, or buy Saft, and add it to Safari!

log_statement = true
syslog = 2
syslog_facility = 'LOCAL0'
syslog_ident = 'postgres'

The database will need to be restarted in order for the changes to have any effect. To view the SQL queries that are being executed, log in as root and issue the following command:

[root@localhost ] tail -f /var/log/messages

Once you have finished the analysis don’t forget to turn off the SQL logging.

The reason I needed to do this was that Postfix was rejecting email for a virtual domain set up on my server. The mapping tables were in the database so I needed to see what SQL query Postfix was sending to PostgreSQL to try and figure out why the lookup was failing. As it happens I figured out the problem, but that’s a post for another day.