Red Hat back-office team to be Big and Blue whether they like it or not IBM-owned subsidiary Red Hat is docking a bunch of its back-office staff, along with the techies that support them, into the mothership. The migration – which some onlookers may be surprised didn't happen sooner – is due to take effect from the start of 2026. Company insiders were told about the move on September 3. According …
What I wanted was a position in a company that had a major UNIX distribution. I thought I had missed the boat in the late '80s, and then there were the rumours of the RS/6000.
When I left the branch of AT&T that I was working for (UNIX support for switch development) for IBM I was accused of selling out, and in truth, I had regarded IBM as an enemy to UNIX (what was the OSF really for if not to muddy the waters). But looking back 35 years, it wasn't the bad move I thought it could be. It was right at the end of the 'old' IBM, suits even in the support centres and devolved management through your line manager, and in general I found my time there OK.
I lasted 6 years, then the money in contracting and the lack of possible job advancement in a technical role (IBM in the UK was really just a marketing company) lured me away. Since then, I have worked with IBM on many projects, and the company is nothing like it's heyday. But even last century, IBM had a reputation of buying and subsuming other companies (Rolm, Tivoli, and Sequent come to mind, but there were many, many more). They all disappear eventually.
Still, getting on the AIX bandwagon secured my working life, and I can probably look forward to the short time before retirement while still working with a UNIX variant.
I thought the same until I worked there early in my career albeit as a contractor. Worked in IBM Hursley which is an amazing place and programming OS/2 was fun. But dear god the corporate culture was stifling - everything was by committee and mystery requirements with layers of management stretching to infinity. It was like entering a parallel universe where the entire software stack was unnecessarily different from everywhere else because IBM ate its own dog food even if the dog food was mouldy and 10 years out of date. I remember having read email on some dumb terminal where you had to refresh the screen to see if email had arrived because the terminal couldn't refresh on its own.
Several things. Hursley is relatively empty now. Oh yes, people are working there, but there is precious little development going on, most of the people there are either supporting things like CICS and storage, although there is still a major datacentre there, but a lot of the space is leased out to other parts of IBM, and even to other companies.
This can be seen in the car parks, where there is lots of empty space, even at peak times.
The second thing, about email. Up until about the '90s, IBM was a mainframe focused company. Their internal systems were based around mainframe and 3270 type terminals. PROFS/NOSS was a capable system, it just worked differently than PC's and UNIX systems. You can say that it was dumb, but the whole model was different. In a mainframe world, the central system did not do async. upgrades, it relied on retentively intelligent terminals to request, buffer, display and transmit back data on a screen-by-screen basis.
Where PCs were deployed, and before the SNA/TCPIP inversion in the '90s, they normally fitted into this environment with 3270 terminal emulator software, and co-ax adapters installed in them connected to 317x or 327x communication controllers.
What this allowed was for you to have hundreds or sometimes even thousands of terminals connected to a mainframe, when the average UNIX/VAX system could maybe manage maybe a hundred or so maximum, and often much less.
Whilst early IBM 3270 terminals looked like they were built to install on battleships, they were quite clever (and expensive!) Comparing to other types of terminal, you could say that the closest contemporary(ish) DEC terminal would have been a VT131 fully kitted out with memory. They could take a form from the host system, with field delimiters and some integrity constraints, allow the user to fill in fields without requiring any host resource, and then transmit it back to the host when ready. This was what gave it a screen-by-screen operating method. They were far from dumb.
To get something similar on a serial terminal, you often ran the terminal in full duplex mode, with the host being involved in almost every keypress. That limits the number of terminals you could have.
You just didn't like that mode of operation because it didn't fit with what you had used before, but for many IBMers, it was their way of working.
But when OS/2 was being written, IBM was already beginning the transition from this. They bought Lotus specifically for Notes/Domino, and then slowly transitioned the entire company across to email on Notes. By the end of the '90s, PROFS/NOSS was gone, except in one or two holdout bits of IBM.
You have to remember that IBM pretty much invented connected office systems. Through 3270 terminals, 317x communication controllers and many, many other devices, IBMers as early as the late 1970s and '80s had access to systems across the world using VTAM and VAMP, and SNA type networking allowed mail and file transfer around the world while UUCP was the most common technology in the UNIX world. It's just that the rest of the world took a different direction.
But I remember talking to many older IBMers who actually found Lotus Notes completely foreign to them, and they regretted the shift. It's horses for courses.
When I started working at the newly created IBM UK AIX support centre in 1990, NOSS, RETAIN and EHONE were systems I used regularly. Although I came from the UNIX world, I made the effort to try to use the tools effectively, and I was far more comfortable than many of my other people in the centre. But that is not to say that for the first couple of weeks I didn't have to restrain myself from throwing the (heavy) keyboard of the 3270 I had on my desk through the screen and walking out!
RETAIN (specifically RETAIN/370) was a system ahead of it's time. Written in the late 1960s and rolled out in 1970, and running on bare-metal mainframes accessed via 3270 comms, it implemented a global distributed problem management system that allowed you to see, pretty much in real-time, what was being reported from around the world. I did not come across anything as effective as it until this century, when such systems started becoming implemented on the Internet. Yes, it was clunky. Yes, it was much more suited to mainframe people. But it was such a good tool for researching workarounds for current problems by being able to see what the rest of the support community were doing for their customers, and identifying solutions for customers calling in. You just had to learn how to use it!
At the time I was using OS/2 as my desktop for development and the virtual terminal ran in a window onto what I guess you mention was 3270 above. I didn't care to find out. All I knew was it was awful. 5 years before I had been using pine on VT100 terminals attached to Unix and it was superior to what IBM made their own engineers use. That's even before considering even back in the day Windows users and even X11 users would have pretty decent graphical email software they could use. But not IBM.
Like I said it was just this weird otherworld of proprietary technologies and stacks which were pretty much dead outside of IBM but forced on the people inside. The only IBM software I actually liked using at the time (other than OS/2) was a version control system called CMVC which was really pretty neat and ahead of its time (sort of like a proto JIRA + VCS). But like everything else in IBM it was virtually unused anywhere else. I had the misfortune to use Clearcase outside of IBM later in my career and I wouldn't be surprised if some poor bastards in IBM are still forced to use it. Likewise with Lotus Notes which made my life a misery in a few places.
Since IBM sold Notes to HCL, they've stopped using it internally. Corporate email is now Outlook/Exchange on the web, and Teams is slowly creeping in to replace Slack etc.
On my current local project that is currently winding down (I hope, everything is so out of date!), I still have to use both Notes and Clearcase/Clearquest. When the project is shut down, all of these will probably go.
So you used Pine on UNIX. As well as Pine, I've used mail on Unix edition 7, the message utility on MTS, Mail on BSD, mailx and elm on SVR3, PROFS/NOSS on VM/CMS, HP OpenMail on Windows, Outlook on Windows, Notes on OS/2, Windows, AIX and Linux, Evolution on Linux, Mozilla and Thunderbird on Linux. I'm sure I've missed some out. I've been using email for decades.
They all work differently, and there is a learning curve between each one. Just because you didn't get on with NOSS or Notes doesn't mean that they were automatically bad.
Lots of people had different experiences, and you tend to like what you are used to.
At uni, fellow student invented a mail system for our CDC mainframe which required zero lines of new code, and zero lines of modified code. The caveat was that his system was purely internal to the mainframe.
Anyone wishing to participate in this system created an empty file named "mail", and flagged it as public and append-only.
Anyone wanting to "send mail" created/edited a file with their message in it, then copied that file to the recipient's "mail" file, which tacked the message onto the end of the recipient's mail file.
We got "real" email when our PDP-11/45 running 6th Edition Unix was installed.
You just didn't like that mode of operation because it didn't fit with what you had used before
3270 was and is terrible, because 3270 is terrible...
There's fields where you have to type, but no visual indicators where they are, except maybe a cursor color-change once you've moved it into the right spot. Arrow/Home/End/etc keys don't jump between input fields or to the start/end of the one you're on... no, 3270 just lets you put your cursor any damn where you please, across the entire display. You're free to stick your cursor any damn place and give typing a try, no problem... It won't work until you find the right spot, but hey, whatever, have a blast.
On a Unix terminal, you can just fly through the data entry... Cursor goes to the spot where you need it, doesn't/can't go to some spot where you can't type anything. Some simple key sequence jumps you between fields, no need to push a button to clear the screen when its full, etc.
Watching an IBM SysProg on a tn3270 session is utterly embarrassing... Having to hit the arrow keys 87 times to get the cursor to the spot they need it to be to type a command, make them all look like hunt and peck typists on their first day on the job.
Normally <Tab> moves you between fields, and <Shift><Tab> moves backwards. This behaviour made it's way into many UIs influenced by IBM's SAA presentation interface. I think that there were also the concept of some a key taking you to the first field, and another to the last field, like Home and End are used on PC keyboards (but the 3270 did not call these <Home> or <End>).
The way that traditional UNIX terminals and terminal emulators work is that every keypress goes up to the system, and up through the comms. stack to the application, is interpreted by the program, and the result sent back down the stack to the terminal, even for just character echoing. Just think, interrupts and potential context switches for every key stroke on every connected terminal session. It may not be much of a concern now, but when computers were uniprocessors and those processors were much less powerful than now, it made sense to offload this from the CPU.
In a 3270 environment, none of this gets beyond the remote terminal controller, even if it gets that far.
Editing a file using something like xedit on a 3270 terminal session looked strange, because it worked on a subset of the file, normally something like 20 lines, at a time. This could look cumbersome to someone used to vi, emacs or any of the other interactive screen editors, but proficient users could work just as fast or faster. I've seen vi and edt (DEC PDP11 editor prior to EVE or TPU) just using the arrow keys and deleting and inserting single characters at a time to manipulate files in a really inefficient manner, when there were better ways of working using search and larger movements. Some people just are not very good at using the tools available to them!
And the 3270 model was aped by other manufacturers, including DEC. The block mode of VT132 and 131 terminals was designed to work in a similar fashion to 3270, and remote terminal controllers were made more efficient by buffering key presses, but very little UNIX software was made to take full advantage of this. Many other system makers had similar terminals and operating models to 3270.
Back in the UNIX world, PDP11's were renowned for their rapid interrupt handling and efficiency, and some of that made it's way into VAXes, but many other UNIX environments were nowhere near that level of efficiency, and spent much of their CPU time just processing character input.
I'm not knocking UNIX (how could I!) it's just that I'm open minded enough to recognise that for the way that they were used, 3270 comms. were just hugely more efficient in system time, and if someone bothered to learn how application software used 3270 type terminals, it didn't really get in the way of the user's efficiency.
But transitioning from one to another in either way was hard.
"We have now micromanagement, decision making from middle management that clearly have no idea of what we do and how and trying to implement ideas that they read somewhere, with no context, data and not giving answer or addressing feedback."
Nevertheless the product remains reassuringly expensive for those manglements who need to tick the "Is product supported?" box or want somebody to sue if it all anything goes wrong.
This should be interesting as slowly absorbing RedHat will change its nature. I'm not sure if that will work out for IBM. In some ways I can see some of the cloud market coming back to RedHat or similar. There will be a recognition eventually that some things are best fully under your own control and Vmware are pissing people off.
Footnote: I was hired at RH in October 2018. Less than a month later the purchase offer was dropped.
The chant from both teams on the purchase was that RH would continue as it was and IBM would have no influence.
It took less than 8 months before the pivot. Jim Whitehurst was appointed President of IBM. And less than 5 months later resigned. I have no need to ask Jim. It was abundantly clear by that point that IBM culture, empolyee policy, management technique was being leveraged on the Redhat structure. I KNOW why he left, leaving behind some very healthy bonuses and such. Very Very Very sad set of circumstances for anyone working at Redhat.
Thus this article was in no way shape or form *news*. It was expected.
The Register 
Biting the hand that feeds IT
Copyright. All rights reserved © 1998–2025
