Technology Tales

Getting an Epson Pefection 4490 Photo scanner going with Ubuntu GNOME Remix 12.10

My Epson Perfection 4490 Photo scanner has been in my possession for a while now and its impossible to justify any replacement given that it both works well and digital photography has taken over from its film predecessor for me. Every time I go installing an operating system afresh, I need to reinstate it again and last year’s installation of Ubuntu GNOME Remix 12.04 only saw me do the deed recently. When I did so, it was brought back to me that I’d never gone and documented on here how this was done. Given that I sometimes use this place as a repository of stuff to which I need to refer again in the future, it seemed remiss of me so here it is for you all.

Though I had XSane and SimpleScan already installed on the system, Sane wasn’t on there so I went and added it and a few other extras using the following command:

sudo apt-get install sane sane-utils libsane-extras

Then, it was onto the Epson website for their Perfection 4490 Photo Linux drivers since Sane’s support for this scanner seemingly remains incomplete even though it pre-dates my move to Linux in 2007. Three files were needed and the following commands install them (depending on when you do this, the file names may be different so just change them to whatever they are for you; it can be done with a single command too but there is not enough girth for that here):

sudo dpkg -i iscan-data_1.22.0-1_all.deb
sudo dpkg -i iscan_2.29.1-5~usb0.1.ltdl7_i386.deb
sudo dpkg -i iscan-plugin-gt-x750_2.1.2-1_i386.deb

With those in place, there was one other task that needed doing so that scanning could be done without resorting to running scanning software using sudo privileges. To free up the access to a normal user account, I needed a HAL device information file. These normally are in /usr/share/hal/fdi/ but they change every time an installation so any modifications that you may make are going to be lost. Therefore, there is no point modifying either /usr/share/hal/fdi/preprobe/10osvendor/20-libsane.fdi or /usr/share/hal/fdi/preprobe/10osvendor/20-libsane-extras.fdi where scanner information usually is to be found.

The first task in creating an fdi file was to issue the lsusb command and look for a line corresponding to my scanner. This is the one that I got:

Bus 001 Device 004: ID 04b8:0119 Seiko Epson Corp. Perfection 4490 Photo

From this, I gleaned the manufacturer ID and model ID as 04b8 and 0119, respectively. These are needed later on. Next I needed to create the hal/fdi/preprobe/ folder structure under /etc since it was there. Then, I created epson4490photo.fdi in the bottom folder of the tree (/etc/hal/fdi/preprobe/epson4490photo.fdi) as follows:

cd /etc/hal/fdi/preprobe/ && sudo touch epson4490photo.fdi

Then, I edited the new file using the following command:

gksu gedit epson4490photo.fdi &

When open, I added in the following text:

<?xml version="1.0" encoding="UTF-8"?>
<deviceinfo version="0.2">
<device>
<match key="info.subsystem" string="usb">
<!-- Epson Perfection 4490 Photo -->
<match key="usb.vendor_id" int="0x04b8">
<match key="usb.product_id" int="0x0119">
<append key="info.capabilities" type="strlist">scanner</append>
<merge key="scanner.access_method" type="string">proprietary</merge>
</match>
</match>
</match>
</device>
</deviceinfo>

It’s all in XML so the place to look is immediately beneath the scanner name comment. The int attributes of the two match elements immediately following the comment line are populated using the information from the lsusb command output with 0x prefixing both the manufacturer and model identifiers. The element with a key attribute of usb.vendor_id is the former and that with a key attribute of usb.product_id is the latter. With epson4490photo.fdi saved, I rebooted the machine to restart HAL and all was as I wanted it to be apart maybe from XSane making complaints that seemed not to be of any actual consequence. With Epson’s Image Scan! and Simple Scan on the PC, there’s no need to be bothered with those messages. Choice is good when you have it, especially when you have expended some effort to get that far.

VMware Workstation in full screen mode hobbles my keyboard

I have recently encountered an odd situation following my recent upgrade to Ubuntu 8.04: when I use VMware Workstation to run Windows XP in full screen mode, the keyboard no longer acts as it should. For instance, Caps Lock and Num Lock keys stop working as does the Shift key. Logging out and back in again is the least that’s needed to set things right but there has to be a better way to fix the problem. I am not saying that it’s limited to the scenario where I saw it happen but it’s still very odd behaviour. If you have a solution, please let me know. Of course, I’ll keep you posted if I find one. In the meantime, I’ll be avoiding full screen mode with VMware as much as I can.

Update 1:

I have done a spot of digging on this one since and gained the impression that there might be a conflict between VMware and the version of X.org Server that comes with Ubuntu. A restorative trick that I have seen suggested is to issue the following command in a terminal, replacing "gb" with your own locale, but I have yet to see if it works:

setxkbmap -rules xorg -layout "gb"

In any case, it looks as if it is not a permanent fix but just a way to keep working without resorting to system restarts, logging off and back on, etc.

Update 2:

I can now verify that the comand quoted above works for me. Of course, it would better to find a permanent fix and even better for the behaviour never to occur at all but any fix is better than none whatsoever.

A perspective on Linux

I have revisited an old website that I used to have online in and around 2000 that has since been retired for a while. One thing that it had in common with this blog was its focus on computer technology. I don’t remember blogging being bandied about as a term back then, but a weblog would have fulfilled the site’s much better. One of the sections of this old world website was dedicated to Linux and UNIX; this was where I collected and shared experienced my experiences of these. These days, unless it is held in some cache somewhere (rather unlikely, I think), the only place that it is found is what I bundled together in a tar.gz file for transfer to Linux. Irony strikes…

Back then, my choice of Linux was SuSE 6.2, followed by 6.4 from PC Plus DVD. It was the first, and only, Linux distro that I bought after exploring a variety of distros from cover-mounted CD’s in books and magazines. While I did like it, it wasn’t enough to tempt me away from Windows. I had issues with hardware and they got in the way of a move. Apart from what some might judge to be clunkiness, there were fewer impediments on the software side.

I am a DIY system builder and there were issues with Linux support of my hardware, particularly my modem. Rather than being in possession of all the electronic wherewithal that a full modem would need, it got the operating system to do some of the work. The trouble was that this locked you into using Windows, hence its Winmodem moniker. Besides this, my Zip drive was vital to me and SuSE didn’t support it out of the box: a kernel recompilation was in order and could involve losing any extensions that SuSE had actually added. Another foible was non-support of a now obsolete UDMA 66 expansion card.

But improvements in hardware support were coming on the scene. Support for printing with CUPS, scanning with SANE and audio with ALSA was coming along nicely and has matured nicely. Apart from cases where vendors refuse to help the Open-Source community and bleeding-edge hardware that needs drivers to recoded according to the demands of GPL, things have come a long, long way.

Software-wise, the only thing holding me from migrating to Linux was my use of Microcal (now OriginLab) Origin, a scientific data visualisation and analysis package that was invaluable for my work. Even then, that could be run using WINE, the Windows API library for Linux. OpenOffice could easily have replaced MS Office for my purposes unless formula editing was a feature outstanding from the specification. GIMP, once I had ascended the learning curve, would have coped with my graphics processing needs. After committing myself to non-visual web development, Bluefish and Quanta+ would have fulfilled my web development needs. Web technologies such as Perl, PHP, Apache and SQL have always been very much at home on Linux, so no issue there. At that stage, experimenting with these was very much in my future. Surprisingly, web browsing wasn’t that strong in Linux then. Mozilla was still in the alpha/beta development phase and needed many rough ends sorting and the dreadful Netscape 4 was in full swing with offerings like nautilus coming on stream. Typography support was another area of development at the time and this fed through into how browsers rendered web pages. Downloading and compiling xfstt did resolve the situation.

These days, I have virtual machines set up for Ubuntu, Fedora Core and Mandriva while openSUSE is another option. I spent Saturday night poking around in Fedora (I know, I should have better things to be doing…) and it feels very slick, a world away from where Red Hat was a decade ago. The same applies to Ubuntu, which is leagues ahead of Debian, on which it is based. With both of these, you get applications for updating the packages in the distribution; not something that you might have seen a few years ago. Support for audio and printing comes straight out of the box. I assume that scanner and digital camera support are the same; they need to be. Fedora includes the virtual machine engine that is Xen. I am intrigued by this but running a VM within a VM does seem peculiar. Nevertheless, if that comes off, it might be that Fedora goes onto my spare PC with Windows loaded onto one or more virtual machines. It’s an intriguing idea and having Fedora installed on a real PC might even allow me to see workspaces changed onscreen as if they were the sides of a cube, very nice. Mandriva also offers the same visual treat but is not a distro that I have been in a lot. The desktop environment may be KDE rather than Gnome as it is in the others, but all the same features are on board. The irony though is that, after starting out my Linux voyage on KDE, I am now more familiar with Gnome these days and, aesthetically speaking, it does look that little better to my eye.

So, would I move to Linux these days? Well, it is supported by a more persuasive case than ever it has been and I would have to say that it is only logistics and the avoidance of upheaval that is stopping me now. If I were to move to Linux, then it would be by reversing the current situation: going from Linux running in a VM on Windows to Windows running in a VM on Linux. Having Windows around would be good for my personal education and ease the upheaval caused by the migration. Then, it would be a matter of watching what hardware gets installed.