• E-mail this page
• |  Print this page
• | 
• | 
• | 

Readers Chime In On The Greatest Software Ever

Our article on the Greatest Software Ever stirred many reader responses, from a variety of sources, including one from a writer who has a picture of himself next to a running Colossus machine--the machine that cracked the Nazi codes--at Bletchley Park, England. Many of the comments add depth and understanding to my selections. Several writers made good arguments contesting some choices, particularly the Morris Worm and Microsoft's Excel spreadsheet. Then there are the corrections. We'll get to them later.

First, consider this on the Morris Worm.

I am one of the people who actually analyzed Mr. Morris's creation back in 1988. It was buggy, used extremely inefficient algorithms (simple linear searches of long tables), and had blocks of dead code. It didn't do what the comments claimed it should, had stolen code integrated into it (Morris didn't write the password cracker—he took it, without authorization, from another Bell Labs project), and above all was used in what was judged to be a federal felony.

I think your characterization of that code as one of the 12 greatest pieces of software written is quite a bit off the mark.

This message was sent by Gene Spafford of the Purdue University Center for Education and Research Information, which researches information security issues. Spafford was an expert witness for the prosecution at the Morris trial.

Now here's a question: Morris was a grad student at Cornell, not a Bell Labs employee, so how did he steal a password cracker from the lab? Whatever the answer, Robert Morris showed us something that could be done that we didn't realize, and he came up with the right combination of code to do it.

Another reader, Raymond Lillard, a control system designer for semiconductor manufacturing equipment, wrote about the worm: "While I recognize the fiendish cleverness of the Morris Worm, cleverness is not enough to make such a list."

Rik Farrow, an editor at usenix.org, a software developer group, writes that the worm exploited a buffer overflow in the BSD Unix utility Finger, not in Sendmail, as our story reported. It exploited the Debug command in Sendmail. Strictly speaking, that's correct. You will find the worm's exploits described different ways in different places, and the Sendmail exploit was an attack that substituted invasive instructions where Sendmail expected a user address to be. In that manner, it was a buffer-overflow-like attack but not a buffer overflow. I read four descriptions of the Morris worm, no two of them alike. I should have read Rik Farrow's. Here's what he's got to say:

The Morris Worm really was amazing software. Its crude algorithms allowed it to spread exponentially and resulted in the Worm infecting 10% of all Internet connected systems in a little over 10 hours, still a record (not for speed, for effectiveness of infection). The use of buffer overflows, the "grappling hook" program that fetched the Worm code, and the multi-pronged approach presaged the development of other malware by many years.

For what it's worth, I still defend the choice of the Morris Worm. It gets top marks in one category: Thinking outside the box.

Bernard Hodson said he had his picture taken in front of the British Colossus machine after World War II, alongside codebreaker Tony Sale of the British M15 cryptographic unit. Hodson wasn't a codebreaker himself; he happened to be at Bletchley Park doing a presentation.

Hodson says Sale challenged any modern computer to break an encrypted message faster than Colossus could in 1944.

But, mainly, Hodson wrote to endorse the OS 360 as a choice for Greatest Software Ever. He said the article took him down memory lane (and I don't mean a row of SIMMS).

"I was one of the first users of the OS 360, having installed a 360/365.... I was one of the very few who negotiated a software penalty clause with IBM, if for any reason the 360 software did not function for our needs," he wrote. Hodson's opinion is the 360 OS was great, but just in case, get your software guarantees written in blood.

Speaking of the 360, one writer, senior consultant James Garrison at consultants the Athens Group, contests the notion that Sun Microsystems co-founder Bill Joy, with his BSD distributions, was one of the first examples of sharing open source code on magnetic tape. "IBM's VM/360 was distributed in source form in the mid/late 1960s and incorporated many contributions from users."

Virtual Machine 360 was a follow-up to the 360 OS. What Jim says may be true, but I don't know of any version of the IBM 360 OS that was treated as open source code. It's not enough that it was distributed on magnetic tape, with user contributions. Were the tapes distributed free to academic institutions, researchers, and the general public?

He goes on to nominate VM/360 as great software:

While OS/360 was a great system, and its genesis has been forever enshrined in the software pantheon by Fred Brooks' masterpiece [The Mythical Man Month], it can be argued that the real unsung star was VM/360. This OS embodied many concepts that have since been reinvented for the PC space [the current wave of virtualization on Intel and AMD hardware, such as VMware and Xen]. There have been rumors every few years that IBM is killing VM, but it just refuses to die.

There was a story (possibly apocryphal) about how in the early '70s the IBM Team working on MVS [a rival mainframe operating system inside IBM] wanted VM killed. During a presentation to upper management about the subject, one upper manager wandered into the developer's area and saw that many were using VM to host their test MVS systems so they could crash without impacting other developers, not to mention not requiring an entire real machine to develop on.

IBM's VM operating system survived the MVS challenge and continues to provide us with our model of hypervisor virtualization today.

Reader Martin Goldberg corrected us on our reference to the Xerox Parc Alto as "the first windowing interface, the first mouse, and first unified graphical user interface."

Goldberg wrote that the first windowing interface and mouse "belongs to Doug Engelbart's NLS computer [an ARPA project] developed in the mid 60s and publicly demonstrated in 1968. It was in fact Doug (with Bill English) that invented the mouse.... You can see the video of the now legendary '68 demo at http://sloan.stanford.edu/mousesite/1968Demo.html.

Goldberg adds, "The Alto did make it to market—though in an upgraded format and under a different name—the Xerox Star in 1981."

I think I should concede that I missed a truly great contributor to the graphical user interface in Douglas Engelbart. But I'll stick with the Apple Mac OS as the first popular implementation of the graphical interface that changed the way we use computers.

« Daily News Podcast for Thursday, August 17 | Main | Diversity Programs Benefit Bottom Line »