Windows expands windows to full screen when dragged to the edge

Note to self

I hate this feature. I think it’s reasonable that it be available, but it’s an annoying default.

You want to arrange windows so you can see everything, you drag one to the side and, bang, it fills the whole screen; the exact opposite of what you want.

The setting is in a kind of place that is logical once you know it is there but would you look there if you didn’t know? I wouldn’t.


Open Control Panel → Ease of Access Center → Make the Mouse easier to use (Win 7) or Change how your mouse works (Win 10).

Select ‘Prevent Windows from being automatically arranged when moved to the edge of the screen’ option.


Unclench teeth.



Detours by Tim Rogers

Text file from v. 1.0
Processed Tue 30 Oct 00:26:01 AEDT 2018 to timbo

Tim Rogers


Fourth Estate, 2018. 338 pages.

First, a caveat; I’ve been listening to Tim Rogers play music since the millennium just gone. I’ve seen him and an assortment of others play in University refectories, various pubs, Casa del Resaca and the Sydney Opera House. I’ve got the CDs, I’ve got a t-shirt somewhere. On the other hand, I refused to go see You Am I go all nostalgia and play Hi Fi Way and Hourly, Daily back to back a few years ago. What does that mean in the context of these comments? It means I want the book to be good. Objectivity … must be in doubt.

Scan of cover.

The cover of Detours by Tim Rogers.


It’s probably useful to begin with what this book is not. It is not an autobiography. It does not begin and the beginning and end, if not the end, then at least the present. It’s a memoir, a self-portrait in words. Sure, it covers his childhood — or at least a few key episodes. It covers where he’s at now and how he feels about it and how he gets through the days. But you won’t find out how You Am I formed. Or, at least, not very directly. We get a few words on the band he formed with his brother. That’s about it. How did he meet Andy Kent? Or Russell Hopkinson or Mark Tunaley or … it’s not here. What was it like working with Lee Ranaldo or Jackie Orszáczky or … it’s not here. Not a hint of a discography. Most of the albums are not even mentioned in passing. So, in short, this not not a book about You Am I. It’s a book about Tim, and about what was and is going on inside him. This is not a negative remark; it’s just a statement of fact. I think one useful thing a review can do is give a sense of the kind of book we’re looking at. I’ll admit I’ve panned the odd book on these pages, but I generally try to give an idea of what you’re in for and, yes, that includes strengths and weaknesses. But if I think a book might be solid but just not for me, I try to say so.

So what what does the book give you? It gives you Tim. What he thinks, what he feels, how he reacts, how he gets by, what’s important to him, and who’s important to him. Even there, it’s often off screen. Clearly his daughter is a focus of his orbit, but he never narrates a major episode with her in it the way he does with footy friends and drinking companions. She’s always tugging at his mind, but he doesn’t render the encounter, just a few words of a phone call or something. Maybe it’s too central to him to share with us, and that’s fair enough.

If you like the idea of sitting down (or standing up, more likely) in a bar and listening to Tim Rogers pour out his thoughts on everything from footy to Loudon Wainwright to CheesyBite and Davey Lane, this is the book for you. If you want evocative thumbnail portraits of his dad or his first girlfriend, this is the book for you. If you’ve really listened to his lyrics and noted his preoccupations, this is the book for you — though you won’t find out what triggered the writing of ‘Purple Sneakers’ or ‘Heavy Heart’ or ‘We Hardly Knew You’. Not in any specific sense. But you will get an idea of his preoccupations and working habits, and that’s what these things flow from. You will read ‘the boy’s angry at the water’ and a few other insights, but what we’re really glimpsing here is Tim Rogers’ inner world brought out. He’s frank about his mental struggles and how he’s dealt — or not — with them. It’s like a case self-study written by an eloquent subject.

Rogers writes with style. He uses words you might expect — peccadilloes, wankers, cirrhosis — but as always he turns up unexpected and (usually) effective metaphor. There’s no hint of a ghost-writer in a book like this — there wouldn’t be enough money in it to pay one anyway — and every page is stamped with the character of the author. At one point he discusses the trashing of hotel rooms, and figures it’s just mean to the cleaners, and why would a rock star want to pick on a cleaner who’s working long hours for low pay? Hell, they leave a tip if they spill a beer in the room. He’s been way down yet he’s kind of famous and he’s rubbed shoulders with greats. It gives him an interesting perspective on, well, the big questions of life, like “where’s the next drink coming from?”

If you’ve followed Tim through the highs of ‘Beautiful Girl’ and the Lows of ‘Obviously’ and ‘Part Time Dads’, then a lot here will not come as a surprise. But there’s a lot to like about this self-portrait, a cubist view that gives facets and leaves gaps and relies on technique as much as content.

Just don’t come in expecting a chronology, discography or even an
index. It’s not that kind of book.


You and I.

Readability in LibreOffice

As of mid-2018, this seems to be lacking. I know it is of debatable value, but some projects demand a readability level.

Now, the document could be saved to docx and imported into Word. I have experimented with gnu style and diction. For that…

(1) I took the file and saved as .txt.

(2) The style program is in the Debian diction package:

$ sudo apt-get install diction

(3) Run it.

$ style filename.txt
readability grades:
Kincaid: 1.0
ARI: 0.9
Coleman-Liau: 3.4
Flesch Index: 100.9/100
Fog Index: 3.0
Lix: 21.8 = below school year 5
SMOG-Grading: 3.0
sentence info:
414 characters
105 words, average length 3.94 characters = 1.16 syllables
14 sentences, average length 7.5 words
7% (1) short sentences (at most 3 words)
0% (0) long sentences (at least 18 words)
12 paragraphs, average length 1.2 sentences
0% (0) questions
21% (3) passive sentences
longest sent 16 wds at sent 1; shortest sent 4 wds at sent 7
word usage:
verb types:
to be (3) auxiliary (0)
types as % of total:
conjunctions 3% (3) pronouns 10% (10) prepositions 11% (12)
nominalizations 0% (0)
sentence beginnings:
pronoun (4) interrogative pronoun (1) article (1)
subordinating conjunction (0) conjunction (1) preposition (0)

(4) Seems to work OK. Can use flags to tune the output as desired, and do it all on the command line if preferred:

$ odt2txt --stdout filename.odt | style

And of course that could be put in a bash script.

$ cat bin/
odt2txt --stdout $1 | style

Is it possible to create a simple macro to call such a script from within LibreOffice Writer? Probably.

But not right now. It would be easy to do this for LaTeX, too, I’d think. Something like dvitty or dvi2tty then style.


Getting rid of unwanted output from Linux commands

Note to self.

When running a command like find, the output often consists of a mix of error messages and actual output, especially when using something like Cygwin, since there are all those funny directories created by Windows. The error messages are unhelpful, or at least not needed, but can easily be removed.

For example:

username@computername /cygdrive/c/Users
$ find . -name "2004*pdf"
find: ‘./administrator’: Permission denied
find: ‘./Bluepackets’: Permission denied
find: ‘./username/AppData/Local/ElevatedDiagnostics’: Permission denied
find: ‘./username/AppData/Local/Microsoft/Windows/INetCache/Low/Content.IE5’: Permission denied
find: ‘./defaultuser0’: Permission denied

Now, this is actually 2 output streams combined — one to stdout, and one to stderr (standard output and standard error).

Here are two other versions of the same command:

$ find . -name "2004*pdf" > /dev/null
find: ‘./administrator’: Permission denied
find: ‘./Bluepackets’: Permission denied
find: ‘./username/AppData/Local/ElevatedDiagnostics’: Permission denied
find: ‘./username/AppData/Local/Microsoft/Windows/INetCache/Low/Content.IE5’: Permission denied
find: ‘./defaultuser0’: Permission denied

$ find . -name "2004*pdf" 2> /dev/null

The first one redirects stdout to /dev/null (ie removes them from the universe with no errors or files created) and leaves the error messages, the second redirects the error messages to /dev/null and leaves the output.

The first can also be written as:

$ find . -name "2004*pdf" 1> /dev/null

which looks more parallel with the error redirection.

The other time this can be useful is when you want to capture the ‘help’ output of some commands. Most commands send the output of

$ command --help

to stdout so you can type something like

$ command --help | more

If you want to read it and

$ command --help > helpfile.txt

to stick it in a file, but a few send the output to stderr, so the ‘2>’is most useful.

More exciting?

The ZMC page from RSC

My ZMC software used to be (might still be) hosted on the computers at ANU ( But I don’t know how long that will last!

Therefore, this is a copy of that page, with the files linked to the RSC for now. If the links don’t work, let me know. Could also try here.


Conventional crystal structure refinement and solution relies on an analysis of the Bragg peaks, the sharp, well-defined scattering in a diffraction pattern. In doing this, the structure can be considered as an array of identical unit cells, and this reduces the ‘solution’ of the crystal structure to the determination of the unit cell contents.

In a disordered material, while the average unit cell does indeed obey the space group symmetry of the crystal structure, such that solving the asymmetric unit within the cell ‘solves’ the structure, this is not the case for a disordered material — or even for what would generally be considered as an ordered material if considering instantaneous molecular and atomic motions.

Even an ordered crystal will show thermally induced vibrations of the atoms. If we could ‘freeze’ the crystal and look at the atoms, we would find that the space group symmetry is not obeyed. It is obeyed on average. In a disordered material it could be that the average occupation of a given site is 50% one type of atom and 50% another, in which case no site obeys the average symmetry.

What does this mean for analysis of diffuse scattering? It means that you can no longer consider all unit cells as identical. It means that you have to now work with an ensemble of unit cells big enough that averaging across it recovers the average and big enough that it can contain a statistically valid population of the local (short-range order) structures present in the crystal.

This means that analysis of diffuse scattering requires different tools from analysis of conventional Bragg scattering. ZMC is an attempt at a program to allow relatively ready implementation of a simulation of diffuse scattering from (particularly but not solely) molecular crystals. Eric Chan has prepared this flyer to show some of the details of the program, and I am very grateful to him. Further, he has developed a tutorial, available here, and it is really worth a look. His webpage is at and contains modified ZMC versions that can do some great stuff, including modulated structures. He has also done some work on a graphical interface! He’s an amazing guy.

Obtaining ZMC

ZMC is alpha software at best, and always will be; nothing about its performance is guaranteed nor promised. The code and binaries are licensed under the Academic Free License version 3.0; details in the files DISCLAIMER.txt and COPYRIGHT.txt (

Here is a bundle of statically-linked binaries for Linux, from April 2016 — the real current version:

Static binaries for Linux.

Here is a bundle of statically-linked binaries for Windows (compiled using g95 on Win 7; I don’t even know if it is 32 or 64 bit), from April 2016 — the real current version, but not much tested:

Static binaries for Win.

Here is the current source code plus a simple Linux compile script — package has transitioned to gfortran, and all the most important binaries seem to work (thanks to Aidan Heerdegen as always):


Refer to the older packages below for documentation, examples, and so on. Please also refer to this article and its associated deposited material, because it provides a complete working simulation.

Please note that this article was specifically published as Open Access in order to allow for anyone to make use of it.

Here are packages, older versions compiled for different systems. Comments about g95 are not current.

Despite my better judgement, I am putting it up here for download. You can download the current 2014 ZMC in the following formats:
Mac OS X

The README file for Windows is here. Be aware that there are many versions of Linux (and OS X is not known for its long term user support or version-to-version compatibility), so your best bet might well be to contact me about the source code. If you can get g95 working (, version 0.92 I suggest) then the compile is pretty trivial (he says hopefully…).

While ZMC is relatively straightforward to use, and a great deal of work has been undertaken to give it a usable (if text file-based) interface, it is not yet ‘production’ software (and never will be). It is written in Fortran90 and uses a module library developed by Dr Aidan Heerdegen, and some components of the system have not been tested on various combinations of hardware and operating system. Currently, As of 2014, binaries exist for g95 on Intel Macintosh (Mac OS X 10.6), Intel Fortran Compiler on Debian stable on Intel i7 (ie, most Intel-compatible) and Windows xp, and that is about it. So far, once the modules are ported, the main program seems to compile without problems, but I am not in a position to make a general distribution.

Further, there is no comprehensive manual, though it comes with an in-progress version of what will one day, time permitting, be a manual (unlikely). The program contains some basic help options, and the distribution packages contain a (simple) simulation that should work out of the box, and the adumbral manual contains some code that might be useful also.

Some things to note: ZMC is essentially a program for displacive relaxation/equilibration of the model crystal. Any occupancy ordering needs to be done externally. The software needs a suite of toolbox programs, which also need to be bundled with any distribution. These include programs for working with crystal geometry, calculating diffraction patterns (DIFFUSE, Butler, B. D. & Welberry, T. R. (1992). J. Appl. Cryst. 25, 391–399), performing any occupancy simulations and so on.


ZMC is used and the method described in the following publications, amongst others. Apologies for the rough layout.

E.J. Chan and D. J. Goossens, ‘A method to perform modulated structure studies using the program ZMC’, J. Appl. Cryst., 2017 50 1834–1843.

This article and its associated deposited material contain a working simulation with occupancy and displacive elements — a good place to start: ‘A Process for Modelling Diffuse Scattering from Disordered Molecular Crystals, Illustrated by Application to Monoclinic 9-Chloro-10-methylanthracene’, Advances in Condensed Matter Physics, 2015 (2015), Article ID 878463, 7 pages, DOI:

D.J.Goossens, A.P.Heerdegen, E.J.Chan and T.R.Welberry, ‘Monte Carlo Modelling of Diffuse Scattering from Single Crystals: The Program ZMC‘, Metallurgical and Materials Transactions A, 42A (2011) 23-31. DOI: 10.1007/s11661-010-0199-1.

D.J.Goossens and T.R.Welberry, ‘Diffuse Scattering from Molecular Crystals’, chapter in Diffuse Scattering and the Fundamental Properties of Materials, edited by Rozaliya I. Barabash, Gene E. Ice and Patrice E.A. Turchi. Momentum Press, 2009. Print ISBN: 978-1-60650-000-2. E-book ISBN: 978-1-60650-002-6

E.J.Chan, T.R.Welberry, D.J.Goossens, A.P.Heerdegen, A.G.Beasley and P.J.Chupas, ‘Single-crystal diffuse scattering studies on polymorphs of molecular crystals. I. The room-temperature polymorphs of the drug benzocaine’, Acta Cryst. B65 (2009) 382-392.

D.J.Goossens, A.G.Beasley, T.R.Welberry, M. J. Gutmann and R.O.Piltz, ‘Neutron diffuse scattering in deuterated para-terphenyl, C18D14‘, J. Phys.: Condens. Matt., 21 (2009) 124204.

D.J.Goossens and M.J.Gutmann, ‘Revealing how interactions lead to ordering in para-terphenyl’, Physical Review Letters, 102 (2009) 015505-1-4.

D.J.Goossens, A.P.Heerdegen, T.R.Welberry and A.G.Beasley, ‘The Molecular Conformation of Ibuprofen, C13H18O2, Through X-ray Diffuse Scattering’, International Journal of Pharmaceutics 343 (2007) 59-68.

D.J.Goossens, T.R.Welberry, A.P. Heerdegen and M.J. Gutmann, ‘Simultaneous Fitting of X-ray and Neutron Diffuse Scattering Data’, Acta Crystallographica A A63 (2007) 30-35.

Updated May 12, 2016

Clearing radio buttons in Word

Note to self.

When editing a Word document that has radio buttons in it — in my case, a checklist — I accidentally clicked on a button. But I want them all to be blank when I give the checklist to a user!

I have 3 buttons — Yes, No and na, and they are mutually exclusive, so I don’t want to use checkboxes.

But I find that once I have checked one radio button, I cannot easily undo it.

Screen shot showing the 'na' button checked.

I’ve checked the ‘na’ button and now I can check any one button but I want all to be blank.

It can be undone, and it’s actually not very difficult.

(1) Go to developer tab. Now, if the developer tab is not visible, go File → Options → Customize Ribbon and under Customize the Ribbon choose Main Tabs and then check Developer.

Screen shot showing the Word Options dialogue and within that the Customize Ribbon screen

Getting the Developer tab.

(2) On the Developer ribbon, choose Design Mode.

(3) Right-click on the troublesome button and choose Properties.

(4) Find Value and change it to False (just delete True and type False).

Screen shot showing the button's properties.

Change ‘True’ to ‘False’.

(5) Exit the Properties panel by clicking on the red X top-right. The button is reset. It will also reset if you hit Enter after typing False.

(6) Close Design Mode. You’ll see that the buttons still work as they did before.

Note: Sometimes you’ll get a very long list of properties and Value won’t be there; look at the top — have you actually clicked on the button? Beside (Name) it should say the button name, in this case, ‘OptionButton61543’. If it says something like ‘ThisDocument’, you’ve clicked on whitespace beside the button, not inside the button. Just try again! Maybe click elsewhere on the page and then go back to the button. Try clicking on the button then right-clicking, etc.


Clear for action.

Virtual machine without administrator rights: QEMU

We’ve got a Windows 10 notebook and an old game that won’t run on this version of Windows. I have install disks for Windows 98, but we do not have administrator access so we cannot install VirtualBox.

QEMU can do this, kind of.

Went to

Downloaded installer. Ran it.

But it wants admin password…

OK, went to Win 10 machine where I do have admin rights and installed there (both machines 64 bit). It went into C:\Program Files\qemu.

Copied the qemu directory onto a USB stick and then onto the machine with no admin rights, just under the user account (C:\Users\username\qemu)

Opened CMD prompt and cd’d into qemu, which is where the executable files live. Ran

C:\Users\username\qemu>qemu-img create -f qcow2 win98.qcow2 1G
Formatting 'win98.qcow2', fmt=qcow2 size=1073741824 cluster_size=65536 lazy_refcounts=off refcount_bits=16

This created a 1 GB hard drive in native QEMU format (qcow2). So far so good. Tried to boot it.

C:\Users\username\qemu>qemu-system-i386 -enable-kvm -cdrom D: -hda win98.qcow2 -boot d

qemu-system-i386: -machine accel=kvm: No accelerator found

OK, the accelerator probably needs admin access to install a driver or something (I’m guessing, and too lazy to find out for sure; for now I just want to see if something will work). But it might run (slowly) without it.

C:\Users\laurence.goossens\qemu>qemu-system-i386 -cdrom D: -hda win98.qcow2 -boot d

Note: this just says to read the CD from the actual D: (could also be an image file) and to boot from D:.

Booted from CD ROM!

Then install — no worries. Step through the usual tedious Windows 98 install procedure. But it is slow. The lack of acceleration is apparent.

Shows a random screen from the install process.

Screen shot of the Windows 98 install dialogue inside the QEMU virtual machine.

Once install done, shutdown and rebooted with:

C:\Users\laurence.goossens\qemu>qemu-system-i386 -cdrom D: -hda win98.qcow2 -boot c

which boots from C: not the cdrom (D:)

Install completed OK.

Machine boots OK and runs but yes is very slow.

Put in game CD

Not found.

Cannot seem to be able to change CDs once VM is running. Need to reboot to change CDs. I don’t know much about QEMU. Closed down and rebooted with game CD in drive. Installed no worries.

Tried to install SciTech display doctor, but it could not find the info it needed, so stuck with 640×480, 16 colours. Tried reboot with -vga cirrus selected on command line, did not help, Windows 98 still just using generic VGA with minimal capabilities.

Booted without CD:

C:\Users\username\qemu>qemu-system-i386 -hda win98.qcow2 -boot c

Conclusion and to do

It works but is very slow, though the host system is not all that powerful. OK for basic board games, but nothing fancy.

Having said that, I have not spent any time trying to optimise it and it looks to me like anything that does not need admin rights would work, and so I suspect the graphics and sound and all that could be sorted out if I wanted to spend the time. There might be something I can do with acceleration, don’t know yet.

The bottom line is that QEMU does give a working VM running on a Windows 10 host without the need for administrator rights, but it is slow. Could be useful for messing around with old files or programs. Once an application is open, it’s not too bad as long as it is not very graphics-intensive.


Solved, kind of.

Repairing a Casiowriter cw-16: Pictures for Sasha

This is a few more details of the repair that I referred to in this post. It was a very tricky operation, and in giving some details on what I did, I did not want to much dismantle my machine in case I could not get it back together… so this is not completely comprehensive… Here we go.

First, we just take off the semitransparent ribbon cover.

Photo of typewriter

Casiowriter cw-16, top view.

Then, take a close look at the ribbon cassette holder, especially the metal plate that sits on top. If you look closely at the metal plate, you can see that there’s a little bent-down tab on the top edge (as oriented in the photo below). It acts like a spring, holding it in place. A small flat head screwdriver and bit bit of pressing and pulling should allow the plate to be popped off.

Photo of the plate the ribbon sits on

The ribbon cassette holder on the Casiowriter cw-16.

When that’s done, it looks something like the picture below. I have left this as a big image, so you can blow it up and try to see some details.

Photo of the gears that drag the ribbon

Details of the problem part.

Now, in the end I pulled off some more parts — I got that flat ribbon cable that passes around the spindle out of the way, and messed around very gingerly. I can’t say that I had to do that — it might be fixable in place. I was lucky enough to find the little metal shaft that the gear runs on also inside the typewriter. I then put the gear on the shaft, put the tiniest bit of glue on the shaft, using the end of a toothpick, and pushed the shaft back into place using a bit of bent metal. The glue was to both hold the shaft in and fix the crack in the bracket, as per the illustration below. When I pushed it in without the glue, the crack meant the shaft did not stay in place and it fell off again.

xFig drawing of the piece

Rough diagram of the bracket and where the crack was

If the shaft is missing, a suitable short screw or something might do the job.

The tensioning spring needs to be noted. Take lots of photos while working, and take care to avoid any bits pinging across the workshop when they come loose…

NOTE: This does not need to be fixed if using thermal paper and no ribbon, and given that the ribbon cannot be easily replaced (no longer being made, as far as I know), you have the option to just get a fax roll from an office supplies place. Just make sure it is full width, not cash register paper…


Good luck.

Pictures for Andy

This post follows a comment made on post here. Here are some photos of wires and connectors related to the switch/display PCB on the front panel of an AlphaServer1200. If you need to add a PCB, the one with the LCD display and the power, halt and reset buttons — this bit:

Photo of the front of the AlphaServer 1200.

The rectangle highlights the LCD display and buttons that come with the PCB in question.

On the PCB, there are a bunch of connectors, but at least one of them allows you to put the plug on either way around. Here is a picture taken of the PCB. It’s a bit cramped because I did not want to disconnect all the plugs and remove it in case I broke something. You know what it’s like. You pull it apart and it never works again…

Photo showing the PCB in place in the case.

View of the display/switch PCB, edge-on, as installed.

The reversible connector is a flat 4-pin one, front centre of photo. It has 4 wires, red, white, green, yellow. The main thing is the red one goes nearest the edge of the board.The red and white run to the interlock on the top of the machine, the green and yellow run to the main board, with the green one going to a pin near the label J2, as shown below:

Photo showing the pins that the green and yellow wires go to.

The J2 pin on the main board of the AlphaServer.

So the pins are near the front edge of the main board. I hope this helps!



Android tablet part 2: chuck it in the bin

Trying to fix faulty Android tablet. In short: Could not.

You may find the links below some help in pursuing your own fixes, but I could not make it work. I know I should not judge Android by these  cheap no-name tablets from generic suppliers, but it’s hard not to.

This follows from this post which describes things that temporarily seemed to fix things only for them to get worse.

Following, I held down ‘volume up’, then plugged into a Linux box and pressed ‘power’ every second for 10 seconds…  nothing. Checked using gparted, and no unmounted volumes or anything. But I have a Windows 7 VM on that Linux box, so booted that up and, lo and behold, I have a device called ‘Onda V972 tablet in flash mode’.

“6. Get the tablet that should be off and charged up, now press volume up button and while still holding volume up connect the tablet to the PC by the USB, now press the power button repeatedly about ten times while counting at a normal pace or once every second (Keep holding volume up) the tablet should go into download mode.”

Googled “Onda V972” firmware

Website has version 4.4.2, where tablet was running 6.0.1… WTF?

Also, the 972 is very different in size form this…must be misidentified, or is the same apart from the screensize. Maybe open the case and take a look at the board?

Anyway, turns out the PNDP60M7BLK small Pendo Pad computer, Android 6.0.1, is A33 Quad Core GC213BA 6CX2 from AllWinner Tech: “the Allwinner A33 has quad-core Cortex-A7 processor and a dual-core Mali-400 MP2 GPU” ( Mobo says ALONG AL-A33-86VM-V2.0 then R09 2016.09.24.

Search on that suggests:

But it is hard to figure out which firmware image to use for this. Might be able to use Android 4.4.2 or something and then update? Maybe? Possibly? Dunno, I have avoided Android up until now. And I was right to.

Here’s some info:

Download A33-MTABL-iNet-1024×600-gslx680_d71-android4.4.2-20150720-test-keys.rar, but it is very hard to find out what the right file is for a given system.

Maybe something here will work:; but who cares any longer?



OK, downloaded

Could not find a match for the firmware.

Eventually found ET-86V2G-A33-V1.8H.rar at some random webpage. Probably won’t work but at least it is A33.

Do I really want to go messing with these files of unknown provenance?

LIveSuitPack wants to run as administrator. I don’t think so.

Look at PhoenixSuit

It wants admin rights as well.

There is no way I am giving random software downloaded from random servers admin rights on my computer.

I wonder if a ReactOS VM would work?

Download and install ReactOS-0.4.9.iso in VirtualBox

New VM

Type Windows 2003 32-bit, Followed my own instructions.

ONDA device did indeed show up in the Devices menu.

Set up shared folders just because I’d downloaded the software and firmware image onto the host and needed to transfer them.

Copied the unarchived firmware image (since ReactOS does not natively handle rar archives) plus the installers for PhoenixSuit and LiveSuitPack onto the ReactOS desktop.

Installed LiveSuitPack — or tried to; “MFC42.DLL not found”.

Try Phoenix. It installed. But tablet could not be connected to the VM. It appears in the Device list but gives an error message when I try to activate it. Made sure it was not turned on in any other VM and tried again.

Nope, it just cannot find the tablet — and complains about firmware file as well.

First, ADB driver installer. Nope, that does not work either.

All more trouble than it is worth.


Chuck it in the bin.