Brian Edminster makes an open secret of his sharing practices. The consultant, developer and advocate for the HP 3000 and open source software has combined those last two items into a new resource this fall. MPE-OpenSource.org is amassing a collection of public, free-use software that can improve 3000 health and longevity, one porting contributor’s portfolio at a time. His first contact with a 3000 was in high school on a Series II, hacking over dialup using an HP2640A terminal. He started programming on a 3000 in 1978 for an HP Value Added Reseller, going on to work with a 3000 graphics software vendor, with MM II Customiser software, computer administering for the UPI news service, developing and managing an Amisys healthcare system, as well as jobs managing the HMS retail apps used in airports — plus diving outside the 3000 to rewrite and rehost to AIX, Informix, Windows Server, SQL Server, and Oracle.
It’s not common to find a 3000 pro of 33 years who still has the gusto that Edminster displays. He’s been a fulltime independent contractor and consultant for the last 10 years, operating Applied Technologies. Some of the work involves using open source tools to extend 3000 ability. He’s a proponent of the idea that HP 3000s can still pass PCI security audits. And he’s also sponsored OpenMPE with contributions, a very rare profile in the community. We emailed our questions to him after his website just got publishing permission from Lars Appel — an open source porting legend who moved Samba to MPE/iX, among other projects.
What prompted you to start a repository of open source software?
Well, at the risk of being flippant – because no one else was doing it. With the demise and only partial reviving of Jazz, much of the free content was dwindling. Yes, Chris Bartram’s www.3k.com site has some open source apps, and so do some others – but many of them cheated and only linked to the software on Jazz. So when Jazz went offline, so did availability of that software.
Also, www.MPE-OpenSource.org focuses on just that: Open source software. I’m working to host software when I can get permission, and link to it (with backup copy kept) in other cases. I’ll host scripts and freeware (code that’s free but not open source), but that’s not the site’s primary focus.
I am still actively looking for contributions. If you haven’t talked to me yet, and you’d like to host software on my site — even if you have no intention of supporting it — drop me an email.
What would you say are the three most useful open source tools for a 3000 site doing its own administration?
One of my short-list projects for MPE-OpenSource is getting MPE/iX clients published for management systems like XYMon. I’m also playing with some Perl scripts that are designed to make managing disk space easier. “Which 3 tools” is not as important as just making sure that MPE/iX can play nice with whatever an enterprise is using to monitor/manage their other systems.
That’s almost funny. If anything, it’s the other way around. Even though the system is nearly legendary in its robustness, I’ve come to the conclusion that the real reason that the 3000 has served so reliably for so long is because the people that manage it are careful. I call them the Belt and Suspenders crowd. Not that there’s anything wrong with that.
What’s the non-3000 technology that you have found most fun to use?
I’d have to say use of PCs as workstations, in spite of the love-hate relationship I have with them. All kidding aside, we’ve come a long way since the days of DOS and Windows 3.. Using workstation-based IDEs (like ProgrammerStudio) have revolutionized programming productivity. Even something as simple as having a terminal emulator with a display longer than 24 lines makes a remarkable difference. Also, use of data analysis tools on the workstation (even tools as simple as Excel) have changed the way we follow the data. It makes identifying trends and patterns in data significantly easier to recognize.
Why did you donate to OpenMPE during its fundraising drives?
Two reasons: Because I could – and this was primarily because of the 3000. MPE/iX’s been Very Very Good to me. And like my website, it’s a way of giving back to the community.
Secondly, because I want this platform to live on, until the ease of administration and robustness in operation commonly found under MPE/iX application systems is so common as to be taken for granted in any other system.
In many ways, the 3000 was quite a bit ahead of its time, and can still lead by example with regard to how robust a system can and should be. I believe in what OpenMPE was trying to do in those efforts, regardless of where it is now, or ends up in the future.
How far would you estimate, in years, an HP 3000 site can take a production system?
There’s a lot of comments along the lines of “Running them until the wheels fall off.” So I guess you’re asking how long it’ll be until that happens. The answer is: it depends, but from a practical perspective, quite a long time. The parts that wear out are mostly the moving parts (disk drives, tape drives, and cooling fans). It’s already getting difficult to find SCSI disk drives under 9GB.
Ultimately, we’ll have to start using interface adapters, and unless someone gets clever and figures out how to partition a large physical disk into multiple smaller logical disks, we’ll end up wasting the majority of the drive’s capacity (the maximum space addressable by MPE/iX version 7.5 on a single drive is 512GB, and under v 6.5 only 300GB). Really, the only thing we need a tape for is using SLT/CSLT tapes, when replacing a system volume set. Backups will become nearly completely store-to-disk or equivalent with FTP or some other transfer method to an external storage mechanism or provider.
The next software gotcha is the limit of the date intrinsics (at the rollover from 2027 to 2028) but I trust we can deal with that in a similar way to how Y2K was addressed. From a practical perspective, I’d venture to guess that the hardware will outlive the business need for some homegrown software systems, unless they continue to grow — and that’s an especially exciting possibility with the new Stromasys emulator coming available.
Something I’ve discovered is software/systems are like sharks. They’ll die if they stop moving. Even when a businesses’s needs are relatively static, technology isn’t. Eventually, there’ll be a ‘better’ way to solve the business need with newer technologies – unless the software integrates new technologies, as appropriate.