Gaat de formatie nu toch echt beginnen?

▼ De situatie tot nu toe: zes partijen zijn bereid te praten over regeringsdeelname. Hiervan zijn er drie onmisbaar voor een meerderheindsregering: VVD, D66 en CDA. Daar moet dan nog ofwel ChristenUnie, ofwel GroenLinks ofwel PvdA bij om aan een meerderheid in de tweede kamer te komen. VVD en CDA willen graag doorgaan met de huidige coalitie (VVD, D66, CDA, CU), maar zijn ook wel bereid om te praten over VVD/D66/CDA/GL of VVD/D66/CDA/PvdA, maar niet een combinatie met zowel GL als PvdA.

GroenLinks en PvdA willen wel praten over VVD/D66/CDA/GL/PvdA, maar niet over een combinatie met maar één van hen twee. D66 heeft een voorkeur voor VVD/D66/CDA/GL/PvdA en wil (liever?) niet met CU. De ChristenUnie wil eigenlijk helemaal niet zo en voelt zich "ongewenst" door D66, maar is bereid te gaan praten als er geen andere meerderheidsopties zijn. Overigens heeft alleen een combinatie van minstens vijf partijen een meerderheid in de eerste kamer.

Na een congres van de PvdA en een bijpraatsessie van GroenLinks gister is duidelijk dat die twee partijen elkaar vasthouden bij de formatie en dat de leden daar achter staan. Dus hiermee is de optie van VVD/D66/CDA en één linkse partij van tafel. Diezelfde avond wisten journalisten te melden dat dit voor VVD en CDA geen verschil maakt en dat zij nog steeds niet met de beide linkse partijen samen verder willen.

De vraag is alleen of ze dat menen, of dat dit alleen een ultieme poging is om D66 en CU onder druk te zetten om nu te gaan praten zonder dat Rutte en/of Hoekstra in het openbaar een definitieve blokkade tegen GroenLinks en PvdA samen hoeven uit te spreken. Daar kleeft namelijk een groot nadeel aan: zodra die twee partijen afgeserveerd zijn zullen ze vol in de oppositiestand gaan. Het vrolijk demissionair doorregeren van Rutte wordt dan lastiger.

Het is niet helemaal uit te sluiten dat D66 en CU nu toch aan tafel willen, maar dat is toch onwaarschijnlijk. Dat hebben ze immers de afgelopen vijf maanden ook niet gedaan. Voor CU is het ook zeer onaantrekkelijk als de optie VVD/D66/CDA met GL en/of PvdA boven de markt blijft hangen.

Dan toch maar hardop GroenLinks plus PvdA blokkeren? Als Rutte dat doet dan legt dat aanzienlijk meer druk op D66 en CU om te gaan praten over het voortzetten van de huidige coalitie. Maar het is wel een enorm risico: als die twee toch niet willen, of de gesprekken beginnen wel maar lopen vast, dan is de enig overblijvende optie een minderheidskabinet. Lijkt mij niet aantrekkelijk voor Rutte, die dan in het parlement naar meerderheden op zoek moet, maar nu toch wel zonder gunfactor.

Het zou ook kunnen dat Rutte wel met GL en PvdA wil, maar Hoekstra dit blokkeert. Hij staat er immers om bekend ongekend hard te onderhandelen. Dan zou VVD/D66/GL/PvdA een optie zijn. Die combinatie heeft 75 zetels in de tweede kamer. Niet voldoende om stemmingen te winnen, maar wel voldoende om moties van wantrouwen tegen te houden. Hoekstra zou het CDA dan weer moeten heropbouwen in de oppositie, waar Hoekstra zich dan naast ex-CDA'er Pieter Omtzigt staande zal moeten houden.

Als alle poten stijf gehouden worden zal het dus een minderheidsregering moeten worden. VVD/D66/CDA hebben bijna een meerderheid, dus getalsmatig is dat een heel redelijke combinatie die vrij makkelijk aan meerderheden zou moeten kunnen komen. Maar dit is zeer onaantrekkelijk voor D66, met twee rechtse coalitiegenoten en ook nog een in meerderheid rechts parlement. Alternatief is VVD/D66. Dat is evenwichtiger qual links-rechtsverhouding. Die combinatie heeft dan wel al snel drie partijen nodig (kleine 20 zetels) voor een meerderheid voor z'n plannen, dus grote kans dat zo'n kabinet vrij weinig voor elkaar krijgt. En wederom onaantrekkelijk voor het CDA die dan in de al dan niet constructieve oppositie terecht komt.

Of misschien een minderheidsregering VVD/CDA. Die hebben samen maar 48 zetels, dus moeten dan bijna 30 zetels aan steun bij elkaar sprokkelen in de tweede kamer. Met alle rechtse partijen, partijtjes en afsplitsingen bij elkaar lukt dat net, maar Geert Wilders heeft natuurlijk zo een motie van wantrouwen ingediend...

Conclusie: de combinatie GL/PvdA blokkeren is hoog spel voor Rutte en Hoekstra. Misschien dat ze daarmee krijgen wat ze willen: het continueren van de huidige coalitie. Maar het risico is enorm: als VVD/D66/CDA/CU niet lukt blijven er alleen slechte opties over.

Daar staat tegenover dat gaan praten met GL en PvdA voornamelijk voordelen heeft: die partijen zullen dan in het parlement vooralsnog hun tanden niet laten zien, zodat het demissionaire kabinet zonder problemen verder kan, en alle andere opties blijven nog open. Als de onderhandelingen met GL en PvdA mislukken is het een stuk moeilijker voor Kaag en Seegers om de combinatie VVD/D66/CDA/CU af te blijven wijzen.

Tot nu toe hebben Rutte en Hoekstra geweigerd hom of kuit te geven. Maar nu GroenLinks en PvdA zo duidelijk hun bereidheid te gaan formeren hebben uitgesproken zal het niet lang duren voordat verdere (openbare) stilte de facto een afwijzing wordt, en zonder zicht op vooruitgang zal zelfs het engelengeduld van informateur Hamer een keer op zijn en zal zij een eindverslag aan de tweede kamer sturen, met daarin wat de verschillende partijen wel en niet willen. Dat zou heel goed komende week al kunnen zijn.

Permalink - posted 2021-08-29

Amiga short takes, juggling CF cards edition

▼ In this latest installment of Amiga short takes:

• Iconverter v1.0 released
• Just use FAT32 for your external CF cards
• Back up to HDF images
• CardBootDisk
• Joystick restrictor plates
• The Deathbed Vigil
• WHDLoad doesn't like network connectivity

Iconverter v1.0 released

Almost two months ago I released version 0.8 of Iconverter, a tool to convert old Amiga icons so they look right on Amigas running a newer OS version. Version 1.0 will now write ColorIcons by default, but can still write NewIcons. It can also read NewIcons and ColorIcons, and thus covert between the two. And write BMP files. (See the Changes section in the documentation for more details.) Available here and on Aminet.

Amiga 3000 boots a lot better with FastFileSystem added to a boot partition

Because I was getting read errors on the CF card that my Amiga 1200 uses as its harddrive, I decided to play musical chairs with all the SD cards. Which resulted in my 3000 not willing to boot even though it had three perfectly good boot partitions on different drives available. (The superkickstart that needs to be loaded from the harddrive makes the 3000's boot process rather complex.)

These issues went away after I used HDToolBox to "add" the 3.0 version of the FastFileSystem to these boot partitions.

Just use FAT32 for your external CF cards

Replacing the CF card is the easy part. The hard part is getting all the data from the old one to the new one. Now I could slightly cheat because I have an IDE-to-CF adapter than can take two CF cards. With that, it's just a question of setting up the right partitions and copying all the data with copy DX0:#? to DY0: all clone.

In the past, I've managed to mount Amiga-formatted CF cards using a PCMCIA adapter. But that's quite a bit of work and often unsuccessful. However, mounting FAT32 formatted cards in the PCMCIA slot is super easy. So after inserting a third, FAT32-formatted card to temporarily hold the data in the PCMCIA slot, now it's time to...

Back up to HDF images

Obviously it's possible to copy all the files to the FAT32-formatted card. And even though the Amiga isn't too picky about special file properties, there's always a chance that something gets lost in translation. Also, copying individual files is relatively slow. It's much faster to just write an HDF image of the partitions on the old card to the temporary card. The TSgui tool will do that.

This worked very well on my Amiga 3000, but not so much on the 1200: systems running Workbench 3.0 said the image "was not a DOS disk". After I got 3.1.4 back up and running again on the 1200, it mounted the images just fine. However, I couldn't make an HDF image of my largest partition at all. It looks like TSgui is limited by the original Amiga 4 GB drive size limit.

Still, it's not a bad idea to create empty HDF images (or download these) and mount those with ImageMount and copy files to the image rather than directly to the FAT-formatted card. Just make sure to check whether they're readable before switching those internal CF cards...

CardBootDisk

So now you have an empty CF card inside the Amiga and a FAT-formatted card with all your data (and, for good measure, ADF files of the Workbench disks and a copy of HDToolBox). Then you turn on the computer and it doesn't see any of that. So I made a CardBootDisk, which is a floppy that has all the drivers to mount FAT-formatted cards in the PCMCIA slot and ImageMounter to mount ADF, ISO and HDF images. So after booting from that floppy you can partition your new internal CF card and start copying data to it. If you forgot anything, just plug the CF card into another machine (in my case my Mac) and copy the missing data on the card and put it back in the 1200.

Download the ADF image of the CardBootDisk. Tip: if you're installing a fresh copy of AmigaOS, first copy all the files from the CardBootDisk to the harddrive and then install the OS over it. You should then still see the CF card after booting from the harddrive.

Joystick restrictor plates

I got an arcade style joystick recently. I'm not entirely happy with it, but I also don't want to write a bad review. So I'm just not going to name any names. Anyway, the actual joystick mechanism is great, much better than my old joysticks from the C64 days or the ones that come with THEC64. (Too bad the THEC64 won't recognize this joystick when used with a perfectly good DB9-to-USB adapter.)

The thing I wanted to mention is the restrictor plate. This one came with a square one with rounded corners. When playing my favorite game, BoulderDäsh, that doesn't work very well. When for instance you move right, but not exactly at a 90 degree angle, but let's say 95 degrees, the restrictor plate makes it really easy to slip down to 120 degrees or so and now you're trying to walk diagonally. Which makes Rockford walk down.

The fix is to rotate the restrictor plate 45 degrees. Now the round parts catch the underside of the joystick as you move left/right/up/down. But in this configuration, it's almost impossible to move diagonally. So for games that need that, this is unworkable. Solution: get a new restrictor plate. Gaming supply shops sell those. I got an octagonal one, which works fine, but perhaps a round one would have been even better.

Restrictor plates. Who knew. Of course with arcade games a joystick would only be used with a single game, so it makes sense that there are different ones that allow for different degrees of freedom. Something to check for when buying joysticks.

The Deathbed Vigil

Recently I rewatched "The Deathbed Vigil and other tales of digital angst", the movie Dave Haynie made of the last days of Commodore. If, unlike me, you didn't get your VHS copy 25 years ago and still have a VCR to play it, you can now also watch the movie on Youtube. Don't miss the "Chicken Lips Blues" at the 55:30 mark.

WHDLoad doesn't like network connectivity

A good way to easily run demos and games, especially floppy-based ones, is with WHDLoad. On the official website they focus on providing "slave" definitions that allow you to import your existing copy of games, but of course helpful individuals around the internet have already imported lots of games... You then just click an icon and you can start the game.

However, when I tried installing WHDLoad last year, I could only get a handful of games to work, most would just show garbage on the screen. But then when I was messing about with CF cards, all of a sudden all the games worked. Turns out that WHDLoad makes some kind of network connection (as it tells me there is a newer version) and then doesn't work properly. Shutting down the network before starting any games fixes that.

Update:

I blocked cgi.whdload.net so WHDLoad can't perform its update check, but that didn't help. After some further checking, it looks like it's a hardware issue with the way PCMCIA network cards work. I normally use the I-Card, which is a wired 10 Mbps Ethernet card. With that one, if I shut down Roadshow, my networking software, WHDLoad games work. Or, keep the network stack running but unplug the Ethernet cable. But as soon as I plugged the cable back in and the Amiga started to receive (broadcast) packets, the game stopped working. So I'm guessing the issue is with interrupts generated when packets arrive.

With the 11 Mbps Wi-Fi card there's obviously no cable to disconnect, and just shutting down the network stack isn't enough: the Prism driver comes with a WirelessManager program that handles connecting to the right network, so you need to shut that down, too. Use the status command to see the process number (for instance, 6) and then type break 6. Interestingly, taking the wireless network down by turning off the wireless access point also works, even though the activity light on the Wi-Fi card still blinks once in a while.

And... then it turns out this is a known bug. RTFM, Read Those Fine Manuals. Weird though that WHDLoad happily uses the network as you start a game and doesn't issue a warning.

Permalink - posted 2021-08-15

(Re)creating new Amigas for the 2020s and beyond

▼ Two forum posts made me think of what it would take to build an Amiga using modern parts, and what should remain the same and what should change about a new Amiga to remain useful in the next decades.

The first is on AmigaLove, a forum that I found recently and that I enjoy participating on: FPGA Custom Chips. The original post is about creating FPGA implementations of the Amiga's custom chips. That way, it's possible to replace the original chips when they fail, or build a new Amiga from scratch.

The other one is on the MiSTer FPGA Forum: Faster CPU for Minimig core?. (Minimig core = Amiga on the MiSTer.)

(An FPGA is a field programmable gate array, or basically a lot of small parts that perform functions used in all kinds of computer chips, and these parts can be connected in different ways so the FPGA can "become" lots of different chips.)

But what should be our goal?

Obviously, some of us, especially those who still have Amigas that we use from time to time, would love to retain access to something as close to an original Amiga as possible. However, then what about the peripherals?

Today, if you bring an Amiga into a home, there's a 99% change that you can't hook it up to a TV or monitor there, and a 99.9% chance you can't exchange any files between the new Amiga and anything else.

So I say: a new Amiga needs to be more like a THEC64 than an actual Commodore 64 (well, actual Amiga). That means: HDMI for video output and the ability to use USB peripherals.

And that's not impossible. I'm typing this on my Amiga 1200 right now, but I'm using an HDMI monitor thanks to the Indivision AGA MK3 adapter and a USB mouse thanks to the MouSTer adapter. And a USB keyboard thanks to the Rapid Road USB adapter, but I readily admit that that's just showing off. (Although the CTRL keyboard types a lot nicer than a 28-year-old Amiga 1200 keyboard.)

What this could look like is just like with the original Amiga generations, two versions: one integrated with the keyboard, and a "big box" model. I think for the integrated one, a case that holds the middle ground between the A600 and A1200 would make sense. Remove the numeric keypad but keep the cursor keys. This is something new, it doesn't have to look exactly like its predecessors. For the big box version I think something that looks like an A3000 but smaller would be great, or maybe even a smaller version of the Amiga 1000 case.

Chipset and floppies

Obviously to run old games, the chipset needs to recreate the A500/A1000/A2000 chips exactly, as seen by Amiga software. So USB peripherals such as mice, joysticks and keyboards must be remapped to recreated versions of the Amiga hardware. I think that also needs to apply to the floppy drive, as so many games and demos had their own floppy drive code. Perhaps the machines should optionally have an actual floppy drive. That way, access to the floppy-based Amiga legacy is guaranteed. But most of the time, an integrated Gotek-like floppy emulator would do the trick.

(And the middle ground: something that physically looks like a floppy drive and accepts floppies and floppy-shaped objects, but then only reads a barcode from the back to mount the right disk image. That would provide the authentic experience without all the issues with actual magnetic storage.)

Ports

On the one hand, it would be great to have all the original ports, but on the other hand, that is more difficult/expensive, for probably relatively little benefit. Let's have a look:

Mouse and joystick ports: it's easy to map USB mice/joysticks to the Amiga hardware. I'm doing that right now. Then again, the hardware is super simple, so perhaps the ability to use the real thing isn't too onerous.

Video: I'd say: just embrace HDMI and leave the past behind. Of course add good scaling and filters and especially handle PAL vs NTSC well. One of the cool things only the Amiga could do is switch video modes on the fly when cycling through screens.

Then again, the MiSTer FPGA platform does support analog video out currently. Could this be an option? But a complete Amiga video connector, genlock-capable and all? Too much to ask, in my opionion. If you really want that, use an original Amiga in good condition.

Audio: one of my complaints about the modern THEC64 is that it only outputs audio over HDMI. That's pretty limiting. Please give us line out for those sweet, sweet MOD files.

External floppy: not sure how difficult this is. But is it useful? For an internal floppy option, a PC floppy drive can be used (the differences are extremely minor and can be accommodated in a recreated chipset), but external Amiga floppy drives are more complex. And one physical floppy drive along with additional emulated ones seems more than adequate.

Serial: there are a few uses for the serial port that are still relevant, I'd say. So having the Amiga think it has a working serial port would be useful, but not sure if having the additional hardware is very necessary. The Amiga's serial port could be mapped to a USB serial port but also be mapped to a network service.

Parallel: well, parallel printers are dead. Not sure if there's any printers left with reasonable Amiga drivers. Intercepting the Amiga's print output and saving it to a file for later display, conversion and then printing to a modern printer makes more sense. But the parallel port was also used for all kinds of interesting stuff, such as the Digiview video digitizer. Not a priority, but on the other hand: just a bunch of GPIO (general purpose IO) pins seems workable.

A1200 hardware

So that's all the ports you'll find on an A500 / A1000 / A2000. But there is more. Next to the A500, the A1200 is the big gaming platform that you'll want to support to let people play AGA games. So:

Harddrive: lots of Amiga software needs a harddrive. And again, I think the right approach is to recreate something that looks like the original Amiga hardware from the software side, but then remaps to modern hardware. I'm thinking an SD card as the harddrive.

PCMCIA slot: used for all kinds of stuff, but mostly to exchange data with other systems on CF cards and for network cards. I would replace this with an (additional) SD card slot. As using CF cards in this slot required extra software anyway, it would be fine to have an extra AmigaOS driver to talk to this SD card slot.

68020: the A1200 uses the 68(EC)020 CPU. So in order to run software made for the A1200, it's necessary to support the 68020.

More advanced hardware

Retargetable graphics (RTG): most non-game software on later AmigaOS versions will happily run on screens powered by RTG graphics cards. These don't have the limitations of the Amiga's chipset. So RTG support that integrates well with the Amiga's native graphics capability would be extremely useful.

68030: the Amiga 3000 uses the Motorola 68030 CPU. The 68030 isn't very different from the 68020, except that is has an MMU. Not a feature that is used a lot by Amiga software, but still useful in some cases. So not a top priority, but nice to have. 68040 and 68060 were a lot faster, but also less compatible. It's probably more useful to make the 68020/68030 implementation faster than to recreate the 68040 or 68060.

FPU: the Amiga 3000 has a 68882 floating point unit, and Amiga 4000s with a 68040 CPU had a built-in FPU. I think in general Amiga software doesn't require an FPU, but there could be situations where not having an FPU creates issues, such as having to use a less capable version of some software.

And an FPU speeds up some software quite a lot. (I think some people doing MiSTer Minimig benchmarks aren't comparing their results to FPU-enabled results.)

Perhaps giving up some accuracy for speed makes sense here, for instance by emulating the FPU (or CPU+FPU) in software rather than use a cycle exact FPGA implementation to run this kind of software.

Last but not least: networking. No original Amiga came with a network card. So recreating old networking hardware doesn't make much sense. However, some way to implement networking functionality does. Not sure if a bsdsocket.library that simply punts network calls to the ARM CPU that runs a Linux variant makes sense, or perhaps something that looks to the Amiga like Ethernet hardware and is then bridged with Wi-Fi. That way, the Amiga can use "real" networking but doesn't have to keep up with changing Wi-Fi (security) standards.

Well, that's my vision of future Amigas. Now the only thing to figure out is what model number they should get. 1600? 5000?

Permalink - posted 2021-08-08