Gary Kildall and Collegial Entrepreneurship (1997) | Hacker News Reader

Gary Kildall and Collegial Entrepreneurship (1997)

drdobbs.com

33 pointsontouchstart14 comments

Got interested in Gary Kidall after reading this

https://news.ycombinator.com/item?id=12220091

A great read, thanks for digging this up.

One thing that struck me was the comment that

>>> Unlike operating systems before and since, CP/M was not the result of years of research by a team of software engineers.

To me, 1997 seems like it was still the tail end of the microcomputer era, when such things were considered possible (and some of the microcomputer-era diversity remained -- that was the era of BeOS, for example). Today, there seems to be active antipathy to solo coders in many professional settings. I feel something has been lost.

> Today, there seems to be active antipathy to solo coders in many professional settings.

Can you give an example or two of this?

Anywhere that focusses on pair programming would be an obvious start. Similarly, watch out for use of the term "cowboy coder" applied to anyone who'd rather work through a problem in private rather than engage in highly collaborative processes.

Both of these terms, in my experience, accompany a mindset that says that single developers don't produce anything of value.

(This isn't to say that collaboration is valuable or that there aren't many problems that need a bunch of coders to efficiently attack, but rather a fear that the industry is forgetting what a strong solo developer can achieve.)

Thanks for that - quick service even on a Sunday!

I'd be the last to denigrate the value of the occasional developer cage-match at the white board, the technical dive-down lunch or water cooler chat, but I've never worked on a project that didn't leave coders well over half their time to work as they see fit. Or that did pair-coding[1] beyond the several weeks of on-boarding.

I've probably been missing out, but what you describe sounds like a recipe for working hard (read: putting in the hours) without getting code written or having strict accountability.

Is this the new thing? Diffusion of accountability?

I'm probably taking what you are saying to a ridiculous limit/extreme, but that sounds like a disaster for introverts. I mean, it's work, not a social club -- how do the introverts get on in such an environment?

[1] The projects I've been involved with have usually embedded test developers to work in tango with the functional code developers, but I suspect that this isn't what you are referring to as "pair programming".

Pair programming, in this context, means two people sharing a workstation. No "single developer" code allowed, at least when it's taken to extremes.

Agree with much of what you say and...

>>> how do the introverts get on in such an environment?

...in my experience, not great.

Out of curiosity, what kind of area do you work in?

> Out of curiosity, what kind of area do you work in?

I started out in semiconductor design (Motorola) in Tokyo (systems engineering - developing MCU firmware for consumer product customers. Sony and Canon were my main customers). This was a long (?) time ago - the mid-to-late 80s into the 90s. In the mid-to-late 90s, I did a startup developing software for electronics payments and banking on the Internet that sold its wares to the largest banks in various countries around the world (and sometimes the country's big telecom vendor if they were feeling feisty enough to take on their banking clients). In the early-to-mid 00's I tried (and often failed) to construct deals around aquaculture, mobile data telecom back-haul and coal-to-whatever conversion projects. This decade, in a addition to kick-starting software development teams for non-technical startup founders, I seem to be settling in around clinical informatics. I'm hankering to try and see if it is now possible to construct a "remote-first" company for enterprise software in healthcare IT.

But I'm old and old fashioned. I tend to bring in people who have substantial development experience (there are always exceptions for brilliant young talent), and I've learned that people like to work like they like to work and to try to enable their productivity in their own style, not inculcate them in the "one true way" I want them to work.

I'm guessing I said something really out-of-touch to trigger the question. What was it? :-)

There are companies that only do pair programming and apparently never let a developer work alone. Pivotal Labs is probably the best-known example:

https://www.google.com/#q=pivotal+labs+pair+programming

There are some interesting comments in that search, both from Pivotal and from other people, some of whom love pair programming and some who can't stand it.

Thanks!

Here are a few exceptions

https://www.sqlite.org

https://github.com/antirez/redis

You are probably not far of in that assessment. The Amiga 1200 and 4000 had been discontinued just the year before.

Windows NT 4.0 had been out for just a year (it was the version that introduced the 9x UI to NT).

On the hardware side we had Pentium and Pro/II shipping, and also the first AMD cpu without any Intel microcode involved (K5).

In the UK, Acorn Computers were just hanging on (and working on one last RISC OS machine -- which sadly never saw the light of day). At least their CPU architecture lives on, which is more than can be said for many of their competitors.

Funny thing about the UK micros is that just about the only one i am familiar with in the Amstrad CPC, because several of my classmates got one in the second hand market.

I think Kildall answered the solo coder question here

https://twitter.com/ontouchstart/status/762366275339321344

(I post the images via a tweet.)

Reply on news.ycombinator.com