The auto.cifs file shown in the howto requires you to create a credentials file for each machine that you want to access. This is handy when the machines you’re accessing have different logins, but here at work all machines authenticate against an Active Directory server and so share the same credentials.

Initially I solved this problem by symlinking to the same credential file, but that’s still an inconvenience when you need to access a new machine (and boy, do we have a lot of them). Finally I did the sensible thing and altered the auto.cifs file to fallback to a master credentials file if it can’t find a machine specific one. Here’s my auto.cifs file:

#!/bin/bash
# This file must be executable to work! chmod 755!
key="$1"
# Note: create a cred file for each windows/Samba-Server in your network
#       which requires password authentification.  The file should contain
#       exactly two lines:
#          username=user
#          password=*****
#       Please don't use blank spaces to separate the equal sign from the
#       user account name or password.
credfile="/etc/auto.smb.$key"
if [ ! -e $credfile ]
then
    credfile="/etc/auto.smb.master"
fi

# Note: Use cifs instead of smbfs:
mountopts="-fstype=cifs,file_mode=0644,dir_mode=0755,uid=19654,gid=124"
smbclientopts=""
for P in /bin /sbin /usr/bin /usr/sbin
do
	if [ -x $P/smbclient ]
        then
                SMBCLIENT=$P/smbclient
                break
        fi
done
[ -x $SMBCLIENT ] || exit 1
if [ -e "$credfile" ]
then
        mountopts=$mountopts",credentials=$credfile"
        smbclientopts="-A "$credfile
else
        smbclientopts="-N"
fi
$SMBCLIENT $smbclientopts -gL $key 2>/dev/null \
   | awk -v key="$key" -v opts="$mountopts" -F'|' -- '
        BEGIN   { ORS=""; first=1 }
	/Disk/  { if (first) { print opts; first=0 };
		  gsub(/ /, "\\ ", $2);
		  	   sub(/\$/, "\\$", $2);
			   	       print " \\\n\t /" $2, "://" key "/" $2 }
        END     { if (!first) print "\n"; else exit 1 }
        '

The important, and trivial, change is this test:

if [ ! -e $credfile ]
then
    credfile="/etc/auto.smb.master"
fi

All it does is check for the existence of the credfile. If it can’t find it, it uses /etc/auto/smb/master instead.

The beauty of this is that I can use one credentials file for the majority of the file shares yet still provide different credentials for the one or two servers I need to access as a different user (such as an admin user).

First a little history; Oracle bought BEA, makers of Java application server Weblogic, a number of years ago. Sun, obviously, has their own opensource application server Glassfish. Oracle isn’t known for their embracing of opensource philosophies so there was a lot of worry that the Sun technologies would get dropped and merely have the good bits integrated into Oracle’s products.

As it turns out, Oracle is doing some of that, but they’re also keeping the Sun products around but targeting them at a lesser departmental level than Oracle’s own offerings.

Sun’s HotSpot virtual machine is getting merged with Oracle’s JRockit vm, so now there will be one “high performance” virtual machine to choose, hopefully better than the separate ones.

JavaFX, Sun’s attempt at competing with Silverlight and Adobe Flex lives on for some reason. And who knows, maybe Oracle can make it an attractive proposition, lord knows Sun couldn’t.

Sun’s IDE NetBeans lives on, along with Glassfish but will be target at departmental applications rather than enterprise ones. Oracle’s JDeveloper and Weblogic will be marketed as the tools of choice for full scale enterprise apps (mainly because Oracle can charge for Weblogic, I would guess).

Thankfully Hudson, Sun’s open source continuous integration and deployment server/tool lives on under Oracle’s ownership. This is particularly good news for me as I’ve been looking at that for improving the Java development and deployment at my employers (I’d have hated to have gone down that route only to find the product is being mothballed).

As for development of Java as a language, Oracle now has main control over the Java SE, EE and ME specifications and they’ve committed to adding more staff and money to developing them, which is good. Although it remains to be seen which directions Oracle will choose to take the language in.

As an aside, I notice that the Sun website has now been Oracle-ised, which means it’ll probably end up being as painful to navigate as Oracle’s. Look forward to URLs changing on a weekly basis and bookmarks to help pages and so on dying regularly!

The OU is pretty heavily geared towards using Microsoft Windows and associated software when studying and submitting assignments, and I’m strictly a Ubuntu kind of guy, so I’m going to be in for a few issues along the way, no doubt. I’m aiming to use Linux alternatives wherever possible, falling back to WINE to run Windows applications if I have to and I’ll document my progress here.

M865 requires you to use project management software to do some of the assignments and activities, and you’re provided with a copy of Sciforma PS8 to do this. Thankfully, this installs and runs fine using WINE, so I’ll be able to use that if none of the open source alternatives work, but I intend to investigate the linux native options such as Planner, Taskjuggler and Openproj.

I have managed to find and install a linux version of the FirstClass messaging software that the OU use for discussion forums. Whilst you can access the forums via the website, the functionality is reduced and it’s not so nice to use. There’s a debian package available, along with a generic linux archive. It’s a simple case of installing that package and running it on Ubuntu 9.10

The first time you run it, you need to click on the advanced link on the login screen to configure the server to use, this will be either oufcnt1.open.ac.uk or oufcnt2.open.ac.uk. Try both if you have authentication problems, you may not have an account on both.

Next up I’m going to investigate the eTMA assignment submission system and see if there’s anything specific I need to do for that.

The camera and the film I found inside it.

Whilst cleaning the camera I opened it up and found that it still had a roll of film inside. The film label stated it was Ilford HP3 Panchromatic, which turns out to be a forerunner to the current HP5 film, first introduced in 1942 before being replaced by HP4 in 1965.

One of the snow drift shots.

I’ve never processed old film before, and at between 40 and 50 years old, this film was pretty damn old. However, google revealed that in similar situations, this type of film had produced decent results. Following the advice of others, I decided to develop the film in ID-11 for double the length of time stated for HP5 Plus (the modern equivalent of this film).

I was worried that the film would be brittle or that the emulsion would have crumbled off, but aside from being very curly and a little tacky, the film loaded and developed just fine. I think that double the time might have been too much, as the film came out very dark, but otherwise it was ok.

The house build in progress.

So, what was on the roll? Well half of it contained pictures of an extensio being built on a house, and the other half were shots of some deep snow drifts and trees. Not hugely exciting but interesting none the less. I’m going to contact the ebay seller and see if they want the negatives, I’ll also email them the hi-res scans that I’ve made.

You can see the whole set of images on my flickr account.

Our standard recruitment process is to shortlist based on CVs, telephone interview the shortlist and then get the best 3 or 4 in for face to face interviews. After the interview I get the candidates to write a simple bit of java code, the problem they’re given is to implement a number guessing game, essentially “I’m thinking of a number between 1 and 10″.  It’s a simple task, particularly when they’re given certain pointers in the problem description, but I’ve found it is a useful indicator of the candidate’s ability. I’m not looking for perfect solutions, rather I’m looking at how they’ve approached the problem and how they’ve thought about the issues involved.

So far this approach has worked well for us and I’ve been very happy with the people we’ve recruited so far, however I do wonder if the technical test is entirely suitable. Is it challenging enough or do we need different levels of technical test depending on the seniority of the role we’re recruiting for? Can technical tests be realistic enough to give a good impression or should we embrace the artificial nature of the setup?

I’m fairly sure there is no right or wrong answer to this and veryone I’ve spoken to has had a different opinion on the subject. Recently though, I found out that even Gordon Ramsey uses a similar trivial test when he takes on a new chef in one of his kitchens. He gets them to make scrambled eggs:

The idea being, that if they can make great scrambled eggs, then it shows that they know what they’re talking about and will be a good chef.

I now like to think of my technical test as equivalent to making great scrambled eggs. If a candidate can come up with a great solution or implementation of the number guessing game, then they should be a great developer. Of course, not everyone likes scrambled eggs…

Now a bit of background, I’m using a Technotrend S-1500 tv capture card to receive free-to-air satellite transmissions here in the UK (aside: I originally chose this card because it came with a CI slot which I could use with a Dragon CAM to get subscription services from Sky – a not supported configuraiton but working according to my research. However with the release of FreeSat, I decided not to bother and stick to the free to air services, including BBC HD and 8 day program guide). This card comes with a little IR eye and a remote control to make interacting with Mythtv easy.

Configuring Lirc took a bit of trial and error because I couldn’t find a configuration file for the remote control. There’s plenty of documentation out there on how to configure Lirc, that assumes you have a more common remote. One of the best guides is Garry Parker’s MythTv guide, which includes a nice tip on ensuring your remote receiver get allocated the same id each time. However, without the lircd.conf file, you need to create one using irrecord. For me it was a case of running this command:

irrecord --driver=dev/input --device=/dev/input/event6 lircd.conf

Follow the onscreen prompts and away you go, using that file in your lirc configuration.

This served me well until the aforementioned system upgrade to Ubuntu Intrepid. My remote suddenly stopped working, except for the arrow keys and the 0-9 numbers. I went back over the configuratiosn with a fine toothed comb but everything seemed to be correct. I thought that maybe the signal from the remote was being interpreted differently, so I decided to try regenerating the configuration file with irrecord but now irrecord was telling me it couldn’t get exclusive access to the hardware, despite me shutting down mythtv and lircd! Something was very wrong…

After a lot of searching, I came across this page that shed light on the problem and gave me a solution. Essentially, X was grabbing control of the input and treating it like a keyboard, which is why the number buttons and arrow buttons were working but nothing else. The solution is to enter your car’d product string into the /usr/share/hal/fdi/preprobe/20thirdparty/lirc.fdi file as an exception. This mean X will ignore the input and let Lirc handle it instead. Annoyingly, my card is based on the saa7134 chipset and there was an exception for those card already in the lirc.fdi file. Unfortunately, my card’s product string is “Budget-CI dvb ir receiver saa7146 (0)” which didn’t match, so I needed to add another one…

Anyway, I now have a working remote control, which is nice, but I can’t help but feel that this is a bit of a backwards step, in that these devices worked before the changes to the HAL system and now, a fairly large list of exceptions needs to be maintained to ensure that cards continue to function as they used to. I’m sure there are good reasons for the changes but it’s one of those things that as an end user I just find irritating :-)