Archive by Author | Darren

Dillo on Debian: weirdness 2

I noted that Dillo often fails to open the page I go to, or renders it as blank or something. A bit more investigation revealed that the search engine choice is key.

This may be well known, but…

…here is my demo.

I open Dillo, click on the magnifying glass and select Duck Duck Go from the list of search engines and search for:

hermes 10 electric typewriter

and get these results:

Screenshot showing search results.

Results of Duck Duck Go search for hermes 10 electric typewriter

So I click on the second entry (Typewriter manuals) and I get this:

Screen shot of result of clicking through from search results.

Empty screen on clicking through from Duck Duck Go search.

Not very useful. OK, try the same search with Google. Find the same website ( and yes — I can click through:

Screenshot of the webpage.

The typewriter manuals webpage.

It also works through Bing.

Now, this behaviour seems to be erratic. I’ve had other pages that I could not click through to from a Google search. I can almost never click through from Duck Duck Go, and Bing seems the most reliable, possibly related to this here. It’s a bit unfortunate, given that Dillo recommends Duck Duck Go by default…

Anyway, just a little note FWIW.



AlphaServer 1200: adding sound

Had an ancient Creative Technologies SoundBlaster in the shed: Model: CT5803.

Perfect vintage for a box from 1998 running Debian from 2012. Specs of card:

Creative Technology Ltd

DP/N 00074HUW

Sound Blaster Audio PCI 64 Dell/Gateway

(1) Turned box off and unplugged power cords.

(2) Already had Debian 5.0.10 installed. Slotted card into a PCI slot, screwed it in tight, closed up the case and booted up.

(3) (Re)installed a bunch of sound stuff — libasound2, alsa-utils, alsa-oss mainly. Used synaptic and just asked for a reinstall of anything alsa, basically. Also chose a couple of sound mixing applications. One of their dependencies might be crucial, not sure.

(4) Rebooted.

During boot saw: Setting up ALSA … Done


(5) Opened ‘Sound’ on the System/Preferences menu and did tests. Yep, works!

(6) Got a screenshot just to put a pic in this post. Sadly, old versions of browsers won’t work with WordPress (old version of JavaScript is probably the issue), so copied the file off the Alpha to my main box using scp over my local network.

Screen shot with no useful content.

Sound works.

Got to say, it was as easy as that. Did need the reboot, but can’t complain. No monkeying about with firmware or anything. Debian roolz!



PDF booklet-y stuff using pdfjam and pdftk

Say you scan an A5 booklet by opening it flat and scanning each pair of pages. You can then print it out in landscape, stapled on the left and you can read the whole booklet in order. But the pages are out of order if you want to make a new saddle-stapled booklet.

So, let’s say I have a PDF like this one:

Scrteenshot of the pdf, showing the arrangement of pages.

Spread from the PDF of the booklet — pages 4 and 5 scanned onto a single landscape A4 or letter paper page.

And I want to rearrange it so that I can make it into a proper (roughly A5-sized) booklet, stapled in the middle rather than along the edge. Well, there might be a tool for this, but …

(1) We open it in gv and find out that it’s 788 wide, half of which is 394. It’s also 598 high.

(2) Use to make 2 PDFs, one of the left half and one of the right half.

$ -t "0 0 394 0" quietriter.pdf quietriter1.pdf
$ -t "394 0 0 0" quietriter.pdf quietriter2.pdf

Looks good. (Hint: Some PDF viewers don’t view the cropped files correctly — if it looks wrong, try a different viewer before messing with the dropping commands).

Page order in quietriter1.pdf is: back cover (24) inside front cover (2) 4 6 8 10 12 14 16 18 20 22
Page order in quietriter2.pdf is: front cover (1) 3 5 7 9 11 13 15 17 19 21 23

(3) Now, for booklet order, the simplest thing to do would be to put these in order (1, 2, 3, …, 24) then use pdfbook (part of pdfjam).

Sounds like a job for pdftk …

First, we’ll put the first page of quietriter1.pdf to the back. From what I can see, this should work:

$ pdftk quietriter1.pdf cat 2-12 1 output quietriter1a.pdf

(This conCATenates the selected page ranges in the order given.)

(4) Then we interleave 1a and 2 using shuffle, which is designed for just this sort of job:

$ pdftk A=quietriter1a.pdf B=quietriter2.pdf shuffle B A output quietriter_inorder.pdf

(5) Then we use pdfbook to reorder into booklet order.

pdfbook quietriter_inorder.pdf

That gives quietriter_inorder-book.pdf.

(6) We print the file double-sided with flip on long edge. (I just printed it from Acrobat, having done the command line manipulation running the PDF tools within Cygwin.)

(7) Looks good! Of course, there are no bleeds, but a quick saddle staple and then trimming with a guillotine and it looks very nifty, and a lot like an original booklet.

photo of the typewriter and manual.

The Remington Rand (Sperry Rand) Letter-Riter (what a terrible name for a typewriter!) with a facsimile of the manual, produced as outlined here.



A little fix on a Hermes 10

So my Hermes 10 had a little problem; the ribbon vibrator was not dropping down after lifting the ribbon into place. That meant my deathless prose was obscured. Given how many errors I make that may not be a bad thing, but still…

Bitmap image of the typewriter.

Hermes 10 electric typewriter with the upper casing removed.


bitmap of the problem part with the part highlighted

Sticky ribbon vibrator indicated by the red arrow.

I tackled it two ways, both at once, so I don’t know which was more important.

First, I just very lightly lubricated the various bits of the mechanism, working from the ribbon position all the way back to the various lever arms and whatnot near the keyboard.

In doing this, I noticed that while there was a spring drawing the mechanism back (and the vibrator down) attached near the top-right of the keyboard, there was no such at the top-left.

bitmap image showing the spring that was already in the machine.

The arrow points to the spring near the top-right of the keyboard. It draws back the ribbon vibrator mechanism, causing the ribbon to drop out of the way.

Now, in most manual typewriters the vibrator is not symmetrical — it’s operated from one side or the other but not both, so such an arrangement would not be that much of a surprise. But on the 10 the mechanism is more symmetrical, with levers and rods on both sides and the ribbon holder moving perfectly vertically on a rail rather than being lifted up by a lever arm from one side. So it seemed reasonable that there should have been a spring on the left, especially since there was a little pin there with nothing to do but hold the end of the absent spring.

bitmap showing the spring

The mismatched spring I added to the mechanism on the top-left.


To cut a boring story a little short, I found a little spring in my parts bin (not matching, but that’s OK) and attached it, and an the problem went away.

The 10 is not a bad machine. It combines the evenness and regularity of an electric machine with the satisfying thwack of typebars on paper. Its gentle hum is not disturbing. It is a heavy beast — portable only in the sense it is not bolted down — and the key tops sit high due to the electrical paraphernalia underneath. But typing on it is effortless and the result is very tidy.

And now, thanks to my fiddling about, the ribbon vibrator gets out of the way and lets me see what’s going on.

Here are some more pictures.

image of the typewriter, without its upper case.

Some more pictures of the topless Hermes 10.

It certainly isn’t the most invisible fix I’ve done, but it works and makes the machine much more user friendly, so worth doing indeed.

10 10 10 10.

***king Articulate360

Just a note that as of June 2018 Articulate Storyline or 360 or whatever it’s called does some weird and dangerous file handling.

[Note: Apparently the sudden loss of your precious precious work can be mitigated by avoiding unusual characters in file names (spaces being considered unusual) and by working on a local drive (“C:” in Windows world) rather than a networked drive. I am currently trying this to see if it helps.]

So … the context … I am working away, modifying an existing presentation. I click somewhere, and suddenly get bounced back several slides and all my changes for the last X minutes get lost — even if I have saved the file!

The first time it happened I thought, right, I better save the file every change. But that does not help! It is like it is reverting my last few changes and then forgetting them so I cannot redo them — sometimes as many as 6 slides of work is lost!

This is just not acceptable. Especially since people have been having similar problems for years.

In the end, I find I have to save regularly (like every minor change unless I want to repeat it) and after any big changes I close down the program completely then reopen it. I also work on local files (ie on my computer, not the server).

So far, that seems to have reduced the issue to manageable proportions.

Also, no nonbreaking spaces. In 2018, no nbsp. Not impressed.

As of 2018 Alt-0160 does not work in Storyline360, whether using number row or numeric keypad or even the Symbol menu where it is called ‘NO-BREAK SPACE’. Pretty underwhelming — though that is not to say they are not taking the issue seriously — they are.

Does not work regardless of Num Lock setting. I have tried inserting Alt-0160 with and without Num Lock and using Insert → Symbol and choosing character 0160 from the character map. Just does not work. I get a space, but it’s a breaking one, I’ve tried inserting a nonbreaking space in a word document and then copying and pasting that into the Articulate document (this works in PowerPoint, for example) and that does not work either.

Pasting from Word using the different options (like Keep Source Formatting) does not work either.

If I paste a nonbreaking space in from Word (or PPT) and then copy and paste it back from Articulate into Word (or PPT), it turns into a regular space. So Articulate is somehow converting it to being the wrong character. FWIW, en rule Alt-0150 works fine.

I installed the program less than a week ago, so it is completely up to date.

Anyway, that’s my experience of this thing — that they want to charge over $1000 for!

Far out.

Thanks anyway.

Printing PowerPoints with notes two (or more) to a page (repeat)

This is just a text-only version of the previous iteration with some new details. Just for the record. I had to work up the documentation, so why not post it?

This is what I did to print out PowerPoint slides two per page including the notes underneath.

There is (as of 2018) no option in the PowerPoint print dialogue for this. Note also that I do not have access to a full version of Adobe Acrobat Professional or whatever they are calling it at the moment.

(1) ‘Designed’ notes text in PPT; something like 12 pt for notes and 9 pt for refs/sources (Myriad Pro, though Univers works well too). Formatted notes consistently. To work with the slide notes in the text block under the slide, used ‘Notes Page’ under View. Can resize and so on and get WYSIWYG while formatting.

(2) Used CutePDF ( to print to PDF file. Went into Advanced settings for CutePDF and made sure paper was A4. Set Print Quality to ‘1200 DPI’, Set TrueTypeFont setting to ‘Download as Softfont’. Under PostScript Options set PostScript Output Options to ‘Optmimize for Portability’ and TrueType Font Download Option to ‘Native TrueType’. Other PDF print options might be acceptable; don’t know. Mac tends to work better with PDF than Windows. Cannot just use ‘Save as PDF’ option from the File menu because that does not give the notes pages.

The Print settings for PowerPoint (rather than the printer) were:
* Color (sic)
* Portrait
* Collated
* Notes Page (Print slides with notes)
* Print all slides (or the range you want)
* Printer: CutePDF
* No header or footer

Then printed and saved the file.

(3) I decided to crop the resulting PDF using a command line tool, because that is highly reproducible — I can get exactly the same cropped box every time.

I used in Cygwin. Cygwin ( is a Unix-like environment for Windows; would probably work seamlessly in a command line terminal on a Mac if the software tools it draws on were already installed, possibly through Macports or similar. is available from, and also from

The command looked like this:

$ -t "105 112 105 112" infile.pdf outfile.pdf

The four numbers are how much to crop from left, top, right, bottom; I worked them out by opening the PDF in GSview (, a Windows-based viewer for PostScript and PDF files that gives coordinates of the pointer, so I could just move my mouse to where I wanted to crop and note down the coordinates.

I know there are graphical cropping tools, but this works and can do lots of pages. Feel free to use whatever tool works for you, but keep in mind that the ability to make exactly the same crop box every time is very useful. An automatic cropping of white space is not ideal since might vary from slide to slide or vary with alterations of the PPT file. It will also fail to remove any page numbers or footers/headers that are present.

(4) I now had a PDF file with the pages, including notes, cropped to have narrow margins. The next step was to put two onto a single landscape page. You can do this by opening the cropped PDF in say Acrobat and then printing to PDF again, after selecting ‘2 pages per sheet’ or equivalent in the print dialogue.

My preference was to use a command line tool. pdfnup (version 2.08 when I used it) is part of pdfjam, which in turn builds on the pdfpages stuff that is part of LaTeX:

$ pdfnup --batch --suffix '2up' outfile.pdf

This takes the ‘outfile’ from the step and makes a new PDF with two pages per page and adds the suffix ‘2up’ to the file name, so the final file is ‘outfile-2up.pdf’

I have noticed that pdfnup gives less white space around the slides (and thus bigger images and text) than printing two per page using Acrobat.

Darren Goossens

29 May 2018

Hermes 10 — portable as a microwave oven

The Hermes 10 is a funky machine from that generation of electric typewriters that were powered versions of manual ones — they still use a basket and typebars, and the paper goes past on the carriage, unlike say a Selectric where the paper sits still and the print head moves.

This one was bought at auction for less than $20 and it came with a case and a dust cover. The case is enormous. It could carry quite a few changes of clothes were it a suitcase. Here it is pictured beneath an Olivetti Lettera 32 in its case; the Hermes is damn near twice as wide and maybe more than that high. It’s also very heavy. And of course it needs a power socket to work, so the case is really for moving it from office to office. You’re not going to use this machine in Starbucks…

Olivetti Lettera 32 sitting on top of a Hermes 10.

It is big and heavy. It hums and smells a little like ozone — probably burning dust. It seems to work pretty well. The ribbon vibrator does not drop down as quickly as it should; I think it was over-oiled at some point (which is to say oiled at all). It seems to be in nice condition. The visible margins work, everything seems present and accounted for.

Hermes 10 from above.

One of the good things about this old design is that they do not take custom ribbon cartridges — they take conventional nylon ribbons on spools. That means unlike, say, to pick a completely random example, a CasioWriter cw-16, there’s no problem with getting ribbons even though the company behind the product is long gone, or at least long-left the industry.

Hermes 10 with the ribbon cover up, showing the conventional spools.

Touch is … odd. A bit like a big calculator. Pretty easy to get used to, though. Hitting return and watching the whole carriage zip back and the paper feed through is pretty nifty.

The arrangement of keys will be familiar to anyone who knows a Hermes 3000, though the backspace is where us computer users expect it to be — top-right rather than top-left as it is in the 3000. Here is the character set — one lonely accented ‘e’.

Character set from my Hermes 10.

Quite useful. Not as many fractions as older machines, but cents, at, pound and dollar.

The x, =, – and _ characters can repeat (for crossing out and drawing lines), though the _ and – cut lines through the paper; possibly an adjustment is needed or the platen is too hard. It’s a small enough issue and not enough to make me do major works on the machine.

Ser. no. 2052425 (, which places it in 1971.

Anyway, a real bit of big iron.

AlphaServer 1200 — flash the firmware

Following on from this post.

Booted up to SRM console but when tried to boot from console it said BIOS CHECKSUM error (or something like that).

(Note: If you are in AlphaBIOS and want to switch to the SRM console, press the Reset button, wait 5 seconds, then press the Halt button.)

OK, looks like BIOS corrupted. If it’s a bad chip then it’s bad, but maybe flashing the BIOS would work…

OK, flash the BIOS on an AlphaServer

Some possibly useful references:

The latter has three 3 1.44MB floppy disks as zip files, which are supposed to contain all the necessary files. It says:

Copy the following files to diskette #1:

Copy the following files to diskette #2:

Copy the following files to diskette #3 if you’ll be updating I/O option firmware:

But … these are V6.0 files and according to the docs, that won’t work (yet)

Reading the documents, ones finds out that can only go to versions higher than 5.3 by going via version 5.3, because at that version the firmware gets reorganised (written to different parts of the non-volatile memory) — so need V5.3 images. But where from? Spent ages looking for them. Could only find:
File: as1200_v5_3.exe 2573 KB 30/11/04 00:00:00
File: as1200_v5_3.sys 2574 KB 30/11/04 00:00:00

But these are too big for a floppy. and even if I put on CD, no instructions on how to apply. I don’t want to brick the thing… Asked on a forum but got told to use a MOP server — some kind of network booting thing. I am way too dumb to make that work.

More googling; found a V5.3 firmware iso at


But it is for the AS4100 not the AS1200. It contains a file called ./AS4X00/RHSRMROM.SYS which is clearly 5.3, and they seem to be very adjacent machines … do I dare use the 4100 file on the 1200? (It would be instead of the TCSRMROM.SYS you can see listed on disk 2 above). says…:

AlphaServer 1200:
The AlphaServer 1200 is AlphaServer 4000 electronics mounted in an AlphaServer 1000 enclosure.

OK; should be compatible. Do it. Create two floppies

-rw-r–r– 1 username username 806 Apr 25 18:58 AS1200CP.TXT
-rw-r–r– 1 username username 920 Apr 25 18:56 AS1200FW.TXT
-rw-r–r– 1 username username 1536 Apr 25 18:57 AS1200FW.SYS
-rw-r–r– 1 username username 1536 Apr 25 18:58 AS1200CP.SYS
-rw-r–r– 1 username username 4096 Apr 25 18:55 TCREADME.SYS
-rw-r–r– 1 username username 590336 Apr 25 18:55 TCARCROM.SYS

-rw-r–r– 1 username username 806 Apr 25 19:04 AS1200CP.TXT
-rw-r–r– 1 username username 920 Apr 25 19:03 AS1200FW.TXT
-rw-r–r– 1 username username 1536 Apr 25 19:03 AS1200FW.SYS
-rw-r–r– 1 username username 1536 Apr 25 19:04 AS1200CP.SYS
-rw-r–r– 1 username username 4096 Apr 25 19:04 TCREADME.SYS
-rw-r–r– 1 username username 996352 Apr 25 19:02 RHSRMROM.SYS

Note RHSRMROM on the second disk, copied from the 4100 iso file.

I need to make the lfu utility read the correct SYS file, so changed the file AS1200CP.TXT to call for the RH instead of the TCSRMROM file.

Ran lfu (booted up the AlphaServer into the SRM console and typed lfu ).

Selected dva0 (floppy).

Asked for as1200cp (the modified txt file).

It loaded the files (shows on screen).

l(ist) showed the SRMROM to be correct version

UPD> update srm*



Then I found the V57.iso firmware and burned to CD and repeated the process using that. In that case, started the Alpha and at the SRM prompt typed:

P00>>> b dka400:

and it booted from the firmware CD and I installed 5.7 version.

Just accepted the CD’s instructions and then:

update * -all

from the lfu prompt provided by the CD. It successfully updated SRMROM and ALPHABIOS.

Then used my original three floppies to update to V6.0 of SRM.

Finally seems to be OK.

Some funny business with the IO update (‘not supported under lfu’), but seems to be OK, so will just suck it and see.

Is now able to boot from CD without errors, allowing install of operating system(s).


Sneaky naughty google

Google, as you may or may not know, it watching you. They are also watching you when you’re signed out.

OK, so, say you want to sign out of Google. You’re in say Gmail.

You know that if you scroll to the bottom of your list of messages you’ll see a tiny tiny tiny little box labelled (somewhat vaguely, probably on purpose) ‘Details’. If you click it, you get details of all your recent sessions on Google platforms (that’s the right-hand side of the picture).

Screenshot showing the process of logging out.

Logging out of all your google sessions.

OK, so you click the ‘Sign out all other web sessions’ button, close the dialogue box and then log out of the Gmail (or whatever) session you’re in. You are now free of Google.


With all your sessions logged out, try going to:

Screenshot of the web page showing the options.

The list of option on the Google privacy control page.


Click on the first entry — Your Google Search History. You’ll see something like this:

Screenshot showing the Search history slider -- Google is collecting search history though you are logged out.

The Search History slider.

So now even though you’re logged out, there are more setting to try to control how your data are harvested. It’s is almost impossible to reduce the amount of data they suck out of you when you go online, but I hope this helps a little. (Note how they encourage you to log in.)


Oogle google doo.

Why Debian.

Having bought the AlphaSever 1200 and got it to boot, there were a couple of things I wanted to do with it.

First, I wanted to put OpenVMS on it, which I’ll talk about elsewhere.

Second, I wanted to put something modern on it if possible.

A look at Distrowatch suggested that I had three options, effectively; OpenBSD, NetBSD and Gentoo Linux. A bit of poking around revealed that OpenBSD would work, but without X windows. Now, that’s a perfectly valid option — the machine could actually work as a server. But for me, as a user who wants to monkey around and use the machine for more desktop-y stuff, that was not what I was after.

Since the SRM console can only boot SCSI drives, and the machine had some unused SCSI drive caddies ready to go, I decided I’d have to spend more than the $1 I had so far. I found a chap on ebay who had a couple of 300GB SCSI hard drives — and even better, HP/Compaq ones, the entity that had absorbed Digital. I didn’t need postage and in the end I scored 2 x 36.4GB and 1 x 300GB drives, perfect since I had 3 unused caddies.

I’ve got to hand it to the whole SCSI thing. No setting jumpers, no master and slave, no two only per socket like IDE drives; I just plugged the drives into the cabinet and when I next booted up SRM found them and gave them identifiers — dkb300, 400 and 500, the latter being the 300GB.

I had three disks. My options are NetBSD and Gentoo. Is there a third option? I could try for Windows NT 4.0, which shows up on ebay sometimes. But I’m not really interested in old versions of Windows … but old versions made me think.

I use Debian on my main desktop machine and on my Netbook. I had read somewhere during my recent researches that Debian supported Alpha up to 5.0 ‘Lenny’. With Debian’s relatively long-term support, the last Lenny update (5.0.10) was as ‘recent’ as 2012. That’s a while back, but it’s hardly ancient.

What’s more, Debian might not be considered especially user-friendly but compared to NetBSD and Gentoo it certainly does hold your hand. After too much time looking at NetBSD and Gentoo, I decided I’d start with Debian. If I could get it to work, get a working xorg.conf, for example, then I might find out some stuff that would be useful with the more hands-on installs.

The installer uses a generic kernel, but there’s an SMP one in the repositories.

So …

(1) Hunted around on the web and eventually found the Debian archive (, oddly enough), and within that CD images for the first 6 of the 27(!) disks that made up the full Debian catalogue for Debian 5.0 (‘Lenny’) on Alpha. Downloaded the first 6 disks using a fast unmetered connection, plus the LXDE/XFCE disk and the KDE disk. Some direct download, some jigdo. Though I had internet on the machine, it is slow and expensive where I am. Also, I don’t need to worry about security updates, since Lenny is frozen, so a static CD repo is fine.

(2) Burned disk 1 and put it into the CD drive and booted into the SRM console. Typed:

P00>>> b dka400:

then at the aboot prompt chose kernel 0.

Soon the familiar Debian ncurses install screens started coming up and filled me with confidence.

(3) Set hostname and decided to install to 36.4GB disk, dkb400, which Debian knows as /dev/sdc.

(4) Created new partition table. Create ext3 root (/) partition of 34.4GB at ‘end of drive’. Then create swap of 1.8GB at end of available space. Then create 180MB ext2 boot partition (/boot) at end of the available 200MB and then in the last 20MB, right at the front of the disk ‘cos we’ve filled from the end forward, 20MB aboot partition.

This structure was chosen because (a) you need a small space at the front to install aboot, the bootloader Linux uses to boot on Alpha (aboot = Alpha boot), and aboot cannot boot an ext3 file system but can boot ext2, Hence the 180MB /boot partition. But I like the journaling of ext3, so I don’t want the whole drive ext2. I did not bother with a separate /home partition.

I considered a complex structure whereby all three OSs (this, Gentoo, NetBSD, should I get them working) would mount the same big /home file system from the 300GB drive, but I’m still experimenting. That could still be possible later, depending on how I chop up the 300GB drive.

Turned on bootable flags on aboot and /boot partitions. Don’t know if must, but did.

(5) Root password added and created my own user account in the usual way.

(6) Scanned in the CDROMs, or at least all the ones I’d bothered to burn — this sets up apt-get to know which disk each package is on, to allow install from CD. I can still set up a mirror for the odd thing that’s not on disk, but I decided not to do that yet. I’d like to keep my internet use down for now.

(7) Selected and installed software — ‘Desktop environment’ and ‘Standard system’ only. Watched it go, reading the files off the (fairly slow) old SCSI CDROM drive. Install needed disks 1 and 2 but no more.

(8) Rebooted:

P00>>> b dkb400:

Booted straight into working X login (gdm)! Amazing, Debian is ama … hang on, if I drag a window around the screen it gets blurred out and repeated in random places … uh oh. I hate debugging xorg.conf …

Look at /etc/X11/xorg.conf — it is a bare bones one, with most stuff done automagically.

(9) In the end, the solution was to go into the xorg.conf and reduce the colour depth. I say that in one line but it took me a while to figure it out.

The old video card (a VGA-plug Matrox Millennium MGA-2064W-R2 from 1995) is actually a pretty good piece of hardware, especially for its time. At a colour depth (DefaultDepth 16 in the relevant Screen section of xorg.conf) of 16 bits, it gives me modes all the way up to a frankly impressive 2048×1152, which is way beyond what is listed in the technical documentation for the card! (1600×1200)

I haven’t tried anything demanding in video terms yet, but all the modes work now. For reference, Debian current (9.4) running on amd64-type hardware and using the DVI-D cable runs at the monitor’s full resolution of 2560×1440. 2048×1152 from a card from last century seems more than I had any right to expect.

(10) In terms of software, most of the usual stuff seems to be there; LaTeX, various window managers, terminals, graphics, sound and video manipulation programs. Brasero, Audacity, xfce, KDE, Gnome (and it’s Gnome 2, which seeing as I use MATE on my more modern machines is completely fine), FLWM and Gimp and so on and so on. Even DOSbox, which emulates an old Intel x86 processor. There’s no OpenOffice, but even if there was it would be version 2.4, and I don’t think Open/LibreOffice had really hit its stride by then. KOffice is there (this is pre-Calligra) and is useful, and so are Abiword and Gnumeric. A lot of the tools I use routinely have hardly changed in 10 years (links, LaTeX, xFig). Probably the age is most apparent in the office software and the web browsers (especially Iceweasel 3.x, which is functional but a security issue). Old JavaScript limits the web things one can do — editing this WordPress, for example, is a problem.

(11) Installed the Debian package containing the SMP kernel, updated the soft links from initrd and vmlinuz to point to the SMP versions, and that was that. Hint: before rebooting, make sure that aboot.conf also has entries that point to the previous, working kernel and initrd in case something goes wrong.


Thinking about the Debian infrastructure — the servers, the mirrors, the fact that all these CD images can be reconstructed all these years later and used on what is not a current architecture, it’s all pretty amazing to me. That kind of thing can only happen when enthusiasts are empowered.

To get it all to work and only needing the one fix — the xorg.conf pixel depth — is remarkable.

The next thing I’ll try is the more modern Debian available through the efforts of the amazing people behind Debian unofficial ports. The key website seems to be this one:

Alpha and omega.