OpenVMS Turns 30 Years Old

Process Software celebrates the 30th anniversary of the release of VMS v1.0, the first version of the HP OpenVMS operating system. We compiled some short stories of our fondest memories of working with the OpenVMS community from various Process employees. The combined experience with OpenVMS from the 6 participants below totals 132 years. We hope that you enjoy our trip down memory lane.

Process Software Timeline

• 1984: Process was founded and built TCPware (the first native TCP/IP stack for OpenVMS)

• 1997: Process acquired the MultiNet product from Cisco Systems and assumed all Cisco support obligations for the product

• 1999: Agreement with Attachmate Corp. to take over the Pathway customer base and migrate them to MultiNet and TCPware

• 2000: Agreement with Sun Microsystems to assume all sales, support and development for the PMDF product family

• 2002: SSH for OpenVMS released

• 2003: PreciseMail Anti-Spam Gateway released

• 2006: VMS Authentication Module released

1. Name: Hunter Goatley

Position at Process: Technical Manager

Number of years worked with OpenVMS: 22

Most memorable moments:

As a student at WKU:

• Diving into the VMS system internals without the benefit of the microfiche listings. My system manager suffered from an inferiority complex. He wouldn't let me look at the listings because he was afraid I'd learn something he didn't know. So I learned it anyway, using SDA to examine the contents of memory and other fun things. I wrote a number of tools then that became VAX SIG tape staples.

• The first time I crashed the system with some of my code. We had a VAX 11/785 with a hardcopy console. I was writing my first kernel-mode program and thought I had it all figured out, so I ran it. In the middle of the day, with something like 100 faculty and students logged in, I knew something wasn't right when the DCL "$" prompt didn't come back immediately. Then I heard the DECwriter console going nuts and my system manager yelling, "HUNTER!!!" Fun times.

At Clyde Digital:

• Expanding my knowledge of VMS internals so that, within a year, I was responsible for virtually all of the kernel-mode code in all of the security products.

At WKU:

• The creation of my VMS freeware archive, including a lot of my own software, that became more-or-less the place to go for VMS freeware.

• Porting many VMS freeware packages to Alpha in the very early days of the system.

• The fight for BLISS on Alpha. When DEC announced that the BLISS compiler would not be made available on Alpha, I spearheaded a campaign to fight that decision. Among many posts on Info-VAX, DECUServe, and anywhere else I could, I wrote a white paper arguing for why the compiler should be made available. I was told by some DEC people that my paper had reached the desk of Robert Palmer, then-CEO of DEC. I never heard from him, but did hear from a number of DEC higher-ups during the campaign. DEC ultimately included BLISS (and a lot of other tools) on the Freeware CDs, in large part because of the campaign I started (so said the DEC people at the time, and with the help of them and many others).

At Process Software:

• Shepherding MultiNet and PMDF from TGV and Innosoft to Process Software.

• Designing and writing PMAS (and the thousands of rules I've written, rewritten, and expanded over the years).

Position at Process: Principal Engineer

Number of years worked with OpenVMS: Since 1979

Most memorable moments:

At DEC:

• Doing process control stuff using DEC Dataway on VAX platforms. It was fascinating to go to customer sites on support calls and seeing this stuff being used to make steel, brew beer, create fiber optic cables, produce artillery shells, and refine petroleum products, just to name a few. Not a large margin of error for these things!

At MCI Telecommunications:

• I lead a development effort that implemented the largest-scale and fastest (at the time) Alpha cluster and Rdb database system for near-realtime phone call tracking. We were able to reach record retrieval rates well in excess of 100,000 per second, in spite of the DEC Alpha and DEC Rdb engineering claims that we would never be able to come close to that (the word "impossible" was used more than once). DEC used our results to advertise Rdb, and I was granted 4 patents on the methodology I developed to do this.

At Process Software:

• I developed the SSH implementation in all of our products, and continue to support it to this day. Translating UNIX-isms into VMS paradigms can be a challenge, particularly since we still support VMS versions that are 20 years old!
  
  
• I spearheaded the porting of most of our products (TCPware, MultiNet, SSH and PMDF, specifically) to the Itanium platform. VMS Engineering did a great job of making the port easy and working with us, but it still was a challenge to translate some of the stuff done on VAX and Alpha to the new platform, particularly since we go so deep into the VMS kernel for some stuff.

3. Name: Richard Whalen

Position at Process: Principal Software Engineer

Number of years working with OpenVMS: 27

Most memorable moments:

at DEC:

• Testing TPU - I was so good at describing the problems that I encountered when attempting to do various things that the developers could often figure out what function to look at from my description. I wrote an entire user interface to MAIL in TPU.

• Getting my first AXP (it was fast) so that I could port the VAX software RAID driver to it.

at Process Software:

• Porting SCP/SFTP to VMS - created a "select" on VMS that looked at terminal and mailbox I/O as well as network I/O.
  
  
• Getting MultiNet to run on Itanium - memory alignment was the biggest challenge.

4. Name: Jeff Schreiber

Position at Process: Principal Software Engineer

Number of years worked with OpenVMS: "Officially" it's been since 1993 when I started at Process, but it took a few weeks before I realized things seemed very familiar. After some brain rummaging, it occurred to me that I had been using VMS in my High School computer classes, so I've actually been using VMS since 1987.

Most memorable moments:

Nothing can really top delaying a month long vacation for a week to deal with a very strange customer problem. We knew there was a small data corruption somewhere, but by the time the effect of that data corruption was noticed, we were far beyond the point where the corruption occurred.

Eventually I decided to stop trying to be subtle, and sat down to code a 'bomb' that checked for the condition to be present. It was actually the same day that Timothy McVeigh was being put to death, and my bomb code became known as "Gas His A**", and I proceeded to pepper the code with GHA references.

I finished my code, we got the customer on the phone, ran a test; within 15 minutes we had the exact location of the problem. An hour later, fixed!

Another great memory was another data corruption issue down in our kernel. We had been trying to solve it for months before I was brought in, on my first real driver/kernel assignment. After a number of "why isn't this a problem" discussions with Geoff and Hunter; a bunch of them that actually turned out to be little bugs on their own, I came across the actual problem while walking through the little bugs with them.

The details are a little too technical for a non-driver engineer to understand... but when my friends wanted to know what I was so proud of, I came up with a way to describe it.

"You're at a party. You put down your beer. You grab it; take a drink, and you die. Why? Kindergarten rules baby, it ain't yours anymore and someone slipped something in it after you put it down!" It was the first, and probably only, time I was able to tell a work story without my friends’ eyes glazing over.

5. Name: Geoff Bryant

Position at Process: Engineering Manager

Number of years worked with OpenVMS: I first worked with VMS in 1981 on an 11/780 with a 2-digit serial number. I became a serious user in 1983.

Most memorable moments:

• I worked on a turnkey CAD system with VAXes (not at DEC) and our own disk, tape, and graphics terminal system. Around 1985, I put together a version of VMS to boot off our disk and support our devices. That first boot and login through the graphics terminal was sweet.

• My first major coding effort in the TCPware network kernel around 1995. It was fun to see it released and to think of the billions of dollars of transactions at the stock exchange that had requested the work, all going through my code.

• A couple of those bugs that took months and deep study of the kernel and drivers to figure out - they were brutal to work on, but fun when found. I especially remember how good it felt to page our customer late one Friday night after Jeff Schreiber and I figured one of those out.

• Getting on a plane going cross-country to participate in an IPv6 bakeoff around1997 and the code was close, but not quite working. When I got there and checked back here for my email, Hiroto had a patch for me and it worked. The next bakeoff with Ravi was much less stressful, but we had it working years before releasing it.

6. Name: Lauren Maschio

Position at Process: Director of Marketing

Most memorable moments:

• Releasing new products on OpenVMS has always been a highlight for me throughout the years. In 2002 we released SSH for OpenVMS, in 2003 we released PreciseMail Anti-Spam Gateway, and in 2006 we released VMS Authentication Module. These products were the result of our customers’ active participation from concept to testing.

• Working with loyal customers. I am always amazed about the high quality feedback we have received over the years on our products. The beta test feedback we receive is especially impressive and useful. We even had a good response to some beta tests during several holidays. I remember receiving feedback from some of our customers on a PreciseMail v1.0 pre-beta preview on Christmas eve.

• Working with dedicated colleagues at Process. As a 9-year veteran, I am a newcomer to this industry. Most of my colleagues have worked at Process and with OpenVMS for many years. We are a cohesive team and it is fun to work with them. My colleagues have also earned the respect of others in the OpenVMS community. I especially enjoy reading Hunter Goatley’s fan mail from all over the world. Legend has it that one fan named their kid after him.