Sense checking my plan for 'moving' from Mach3.

I was trying to be balanced and diplomatic.

quite a lot of new linuxcnc users who go down the parallel port route seem to have issues (just an observation). Hence I'd suggest that they spent a little money and got some hardware to remove the base servo loop requirement, or at least unload it onto the hardware. As they would probably (in my opinion) have a better + easier experiance.

I'm a bit more direct than robert was.
Quite frankly the parallel port is obsolete today.
You can't readilly find a new PC for sale  that still includes one.
Thats not really surprising. I have spent a life time working with digital printing devices. Even at home I have not used a Parallel port printer for over 20 years and never used a USB port printer. Everything is networked. PC's don't need a parallel port for anything but a small group trying to run CNC machines. BUt there are not enough for the manufacturers to notice.

Sure stick to old and obsolete second hand PC's that still have one if you must but I have not had such a PC laying around for years.

If you want to purchase a new PC even if you bought one with an expansion bay so you could install a parallel port, you would likely come unstuck because  of lack of other network driver support or other issues. Plus you will probably still struggle to get adequate latency to run the parallel port.

Why not just use a USFF fanless industrial PC and some network connected hardware or even a raspberry Pi and get with the 2020's?
My point in mentioning ethercat is that Linuxcnc opens up many options that a new user may not have considered. And I did say it was not a trivial solution.

I would go as so far as to suggest that new users should be told to look for an interface board as it will make their lives easier (and their linuxcnc experience).

I don't want to push just the Mesa offerings (although that's all I've got [7i92 + 7i76, 7i76e & a few SpinX1's), as I know there are other solutions about, but it may be worthwhile maybe having a hardware guide page listing (supported! + current) hardware interface options.

I guess we will always get the user(s) who like to be spoon fed the easy solution and lead by the hand.... and from a perspective of time vampires they are probably best directed to using other cnc offerings (like uccnc / mach3/4) where they buy a piece of hardware and then expect 24/7 support for their purchase, and ask the daftest of questions on the forums because they cannot be bothered to at least glance at the manual (or know where it is even!)

Not to reference Mach4 as a comparison.... but Mach4 offers the LPT port as a legacy compatibility option and it is not pushed (in fact users have to pay more ($25USD) to use it and it does not offer full functionality). www.machsupport.com/shop/parallel-port-legacy-plugin-for-mach4/ The point that I'm trying to make is that for a NEW user experience to linux and linuxcnc should the use of the parallel port should not be pushed and maybe alternative options that get them up and running quicker (with less reliance on latency).

I've never used the parallel port (as I've not got a PC which has one any more or one that would take an ISA, PCI or PCI-E board to even try it) but the route seems to be to:

LinuxCNC 2.8.0 Debian 7 Wheezy RTAI
Users requiring a known-stable RTAI installation can install the Debian 7 Wheezy ISO and then upgrade as described in 2.8 documents.

LinuxCNC 2.8.2 Debian 10 Buster RTAI
The more adventurous can install the Buster ISO and then install the experimental RTAI kernel as described in 2.8 documents

it sounds like 2.8.2 with the RTAI is not stable .... or at least an "easy" solution for new users. All to save $100 usd or so...

linuxcnc can use an LPT port.... but would a new user REALLY want to use it and REALLY accept that it may not give them the user experience (and learning curve) that they may expect. Only for them to get annoyed, go elsewhere and spout about how poor they perceive LinuxCNC is.... and dump $500AUD on another hardware solution, all because they chose the most difficult and "FREE" route.

There are discussions about 2.9 release floating about and being released, and how would this leave the use of the parallel port with RTAI, looking at:
www.rtai.org/?Homepage&id=41
RTAI 5.3 released
May 19, 2021
The tarball is available at rtai-5.3.tar.bz2
Supports up to Linux 4.19.

how does this stack up with bullseye / bookworm?

at some point it's probably going to become legacy with less and less support due to the availability of integrated hardware, and although I appreciate that there will probably always need to be an ultra low cost need / solution, that is probably NOT going to be best to highlight to the majority of new users. Sure it can be done, but it probably wants highlighting that it won't be the easiest path and the one with the best user experience.

robertspark wrote:

There are discussions about 2.9 release floating about and being released, and how would this leave the use of the parallel port with RTAI, looking at:
www.rtai.org/?Homepage&id=41
RTAI 5.3 released
May 19, 2021
The tarball is available at rtai-5.3.tar.bz2
Supports up to Linux 4.19.

how does this stack up with bullseye / bookworm?

at some point it's probably going to become legacy 

Bullseye runs the 5.10 kernel and Bookworm is now on version 5.16. I think Bookworm will be released later this calendar year. I think its fair to say RTAI is obsolete now being 99 versions behind the current development linux kernel which is 5.18

There is now so much hardware that requires Kernel 5.10 and above based on my experience.

the Debian release cycle is about every two years so bookworm will be next year around August in my opinion

​​​​​
wiki.debian.org/DebianReleases#Production_Releases

the Linux kernel releases don't work as a batch of 100, so rtai is not that far behind
en.m.wikipedia.org/wiki/Linux_kernel_version_history

it's just that it is still about 1 version behind the current Debian stable (buster)

4.19 is also an LTS (long time support release) and is going to be maintained to Dec 2024 

5.16 is NOT an LTS version so it probably won't be in the stable release of bookworm as they will choose an LTS version as the stable so that the release lifecycle is extended as long as possible 

bookworm presently uses 5.16 not 5.18, but that is not really important as it seems to use 5.16.18-1 which is probably where you've seen the  "18" from
packages.debian.org/bookworm/kernel/

looking at the LTS kernel versions, I am guessing that bookworm will probably be released as stable with something like a 5.19 kernel?   (can't see bookworm dropping back to 5.15 as the stable kernel if it's being presently tested on 5.16..... so it'll probably use the next LTS release.

LTS releases: 4.1; 4.4; 4.9; 4.14; 4.19; 5.4; 5.10; 5.15...... ?? 5.19 maybe?  (it also seems that they choose the LTS version from around Oct - Dec released kernel version.... hence it would tie in with next years Bookworm stable release of around August time, choosing the next LTS kernel later in this year).

5.10 is a LTS release so its supported until December 2026 + it's also the debian bulleye stable release kernel, hence why manufacturers probably are developing / maintaining firmware support for their devices based upon it.

robertspark wrote:

it's just that it is still about 1 version behind the current Debian stable (buster)

 

Sorry Debian Stable is Bullseye wiki.debian.org/DebianReleases
Buster is oldstable (with the emphasis on Old!)
Bookworm has advanced its kernel since I have been playing with it. From memory it was on 5.14.
I don't believe that Debian will go back a kernel version because of its development cycle.

robertspark post=240781 userid=23480

bookworm presently uses 5.16 not 5.18, but that is not really important as it seems to use 5.16.18-1 which is probably where you've seen the  "18" from
packages.debian.org/bookworm/kernel/

If you reread what I read, I said Bookworm was on 5.16 but the current unstable pointy end of the kernel (not in debian) is 5.18
If you look at kernel.org's ftp site, there are many more versions than what you think. I thought they were on a fortnightly release cycle. Here is V 5 alone I'll leave you count the changelogs.
mirrors.edge.kernel.org/pub/linux/kernel/v5.x/

So RTAI is well and truly off the pace and teetering into oblivion.

Agree about the Bookworm release. The freeze time line is here
release.debian.org/testing/freeze_policy.html

But this is all academic. Its the availability of kernel drivers that still work for old hardware that will drive obselesence.
I blew away a copy of Mint today on an Office PC because for some unfathomable reason a barcode scanner stopped working on the PC.
In its place I installed Bullseye 10.3 and it was the sweetest smoothest install of Debian I have ever done. The wired Network just worked. Wifi just worked. It found and installed both my network  printers leaving me to quickly add a couple of networked Zebra label printers which was semi-automatic once I gave it IP addresses. Note not a USB or parallel port printer in sight in my world! And my barcode scanner works again!

Fixing auto login, and the screen locker, just left me to install chrome and through the magic of cloud computing the PC was fully operational again!