Sex, Drugs & Unix

Tuesday, November 29th

20 years of USENET


From what I can tell, my first post to USENET was March 20, 1986

Thats kinda cool, in another 4 months it will be 20 years since then.

Russ Nelson beat me by three years, but then, Russ is four years older than I am, and I am not a racist. (UPDATE: This was wrong, and I've issued a retraction.)

I don't find ESR on USENET until May 10, 1987, but its entirely possible that I don't know the right email addresses to search farther back.

I wonder what ever happened to Rex Jollif and Joel Axe?

Jim on 11.29.05 @ 12:05 PM PST [link]


This bothers me on several levels




Jim on 11.29.05 @ 08:31 AM PST [link]


more on esr and GNU


Tellingly there are *no* posts by Eric Raymond in net.emacs.

None. Google's record of the group starts Dec 26 1982, 9:02 pm. The last post to the group is dated Nov 8 1986, 9:52 pm.

Not a single post by Eric "I started Project GNU" Raymond.
Jim on 11.29.05 @ 07:52 AM PST [link]


HTML O' the Day


Well, that's the kiss of death.

... which has been a "springboard" for many careers in pharmaceutical sales.

Well, after 55 hours of pain, cold, and suffering, we have got our Xbox 360 and successfully smashed it outside Best Buy minutes later, with a large sledgehammer in front of a few Xbox Fanboys who just bought brand new 360s.

"Flipping the dynamic".

Irish dirty dancing

Vodka and orange juice is the traditional substitute.

How your life can serve as a warning to others.

Red Hot Papa

Face-down people.

Rule #2: don't deep-fat fry.

Butterball Help-Line Help-Line.

Jim on 11.29.05 @ 07:24 AM PST [link]


more questionable history from esr about the origins of GNU


Here's another ESR recollection that shows on trivial investigation to be questionable:

Richard drew much more from that culture than he seems to remember nowadays. I myself was the person who first suggested to him, at a Boskone in 1983, that Emacs ought to be the GNU project's first product. We'd been friends for nearly five years by that time. It is utterly characteristic of RMS that he says he doesn't remember that conversation, but believes my report that it happened.

As has been shown, the "GNU project" didn't exist until 1984, and wasn't announced until September 1983. Gosmacs went commercial (was sold to Unipress) in April of 1983.

RMS responds on April 19, 1983:

Each time the commercial marketing of Unix Emacs causes someone
to forego it due to price, or to be unable to use it as he would
have liked to due to license restrictions, society as a whole
has been sabotaged a certain amount.

Gosling does deserve a considerable reward for having written
a useful program, but sadly he deserves to lose a lot of that
as penalty for sabotaging its use now that it is written.

Don't let yourself be sabotaged!

Meanwhile, if you are thinking you may be stuck with paying these
prices, and you don't belive in doing something illegal even
if it is good for the world, you still have an alternative.
An editor is being written in NIL. It's at an early stage
but it's far enough along for its implementor to use it to
edit as he adds to it. A Unix that can support shared programs
is coming from Berkeley. NIL for Unix is being worked on
(and for VMS is already available, and public). Since this
will be a true Emacs rather than a semi-ersatz one, it will be
far better than Goslings.

This editor is supposed to be publicly available.
So just hold on a while -- help is on the way.
Sooner if you can help with the work.


Gosling responds to rms on April 25, 1983.

Satotage? I contend that there are more people who don't use Emacs
now because of it's present distribution mechanism than won't because
of it's price. The total lack of support and maintenance has turned
away many people. I get sent tapes and it takes literally months for
me to return them. That turns off far more people than price.
Universites are unusual cases.
Besides. Prices are made of rubber, they can easily change.

On April 28, 1983, Unipress raised the price of Emacs to $1000 for binaries, and $7000 for source code. Unipress eventually backed the price down to $1000 for source.

But back to esr.

Boskone happens in February, typically around the middle of the month, and did so in 1983 and 1984. Boskone #20 happened February 18-20, 1983. Boskone #21 happened February 17-19, 1984.

By March of 1985, we have the first release of GNU Emacs (the one with gosmacs code inside, the very incident which forced the GPL.)

Quoting http://advogato.org/article/512.html:

In 1982, James Gosling wrote the first C-based Emacs called Gosmacs. Initially, Gosmacs was Free Software, but Gosling later sold it to a company called UniPress. However, Fen Labalme, a contributor to Gosmacs, told RMS that Gosling had given him special permission to distribute a free version, so RMS incorporated some of that code into GNU Emacs. However, when UniPress heard about this, they told RMS to stop distributing GNU Emacs. They denied that Fen had been given permission, and even worse, Fen had lost the message from Gosling that could have proved otherwise. Thus, RMS had to comply, but he wasn't happy about it. You have to realize that RMS had been working on Emacs since 1975, and then here comes this company out of nowhere, telling him to stop.

(It strikes me that if Fen was a contributor to Gosmacs, then UniPress had an incomplete copyright to begin with.)

If esr's recollection is accurate, then we must believe that he told rms that Emacs should be the first GNU project while rms was still hacking on lisp machines at the AI lab, some seven months prior to rms annoucing "Project GNU" in September of 1983.

If instead ESR's recollection is off by a year, and the meeting happened in 1984, then we must believe that rms walked away from his newly-enabled life's work after only six weeks, (having quit his job at the AI lab on Jan 5, 1984), in order to attend a science fiction book convention. Even if the meeting did happen, no change in direction was forthcoming until the following September. O'Reilly's "Free as in Freedom" claims that work on GNU emacs didn't start until September of 1984:

In September of 1984, Stallman shelved compiler development for the near term and began searching for lower-lying fruit. He began development of a GNU version of Emacs, the program he himself had been supervising for a decade.

I find the possibility of the meeting and influence that esr claims in 1983 unlikely in the extreme, even in 1984, but of course only esr and rms can tell us if either happened. Eric claims 1983, and this requires that Eric understood "Project GNU" well before rms announced it.

According to Eric, rms says he has no recollection of the event.

Eric's claim is too convenient for me. My assumption, given the above, is that the events Eric claims never happened.
Jim on 11.29.05 @ 06:12 AM PST [link]


More fun with google


Once upon a time, Dennis Richie posted something about character strings in C.

Given the explicit use of character arrays, and explicit pointers to
sequences of characters, the conventional use of a terminating
marker is hard to avoid. The history of this convention and
of the general array scheme had little to do with the PDP-11; it
was inherited from BCPL and B.

Robert Firth posted a correction

A correction here: the C scheme was NOT inherited from BCPL.
BCPL strings are not confused with character arrays; their
implemetation is not normally visible to the programmer, and
their semantics are respectably robust.

Now, Firth obtained his PhD at Cambridge in 1969. It is highly likely that he was actually quite familiar with the BCPL language. BCPL was implemented by Martin Richards, a nearly infamous Cambridge hacker, while taking sabbatical at MIT in 1967, having been previously designed in 1965. BCPL is the language upon which Thompson's B and consequently C were built. BCPL was transported to the Bell Labs by another programming enthusiast, Rudd Canaday. C wasn't invented until 1972.

ESR, in response flamed the snot out of Firth only four minutes later:

I've seen bonehead idiocy on the net before, but this tops it all -- this takes
the cut-glass flyswatter. Mr. Firth, do you *read* what you're replying to
before you pontificate? Didn't the name `Dennis Ritchie' register in whatever
soggy lump of excrement you're using as a central nervous system? Do you
realize that the person you just incorrectly `corrected' on a point of C's
intellectual antecedents is the *inventor of C himself*!?!

dmr countered :

Robert Firth justifiably corrects my misstatement about
BCPL strings; they were indeed counted. I evidently edited
my memory.

Forcing ESR to apologise for his flame:

It is I who have learned the lesson this time. I'd had a rough and frustrating
day, and what I *thought* I saw was some random presuming to know better than
dmr about his own thought processes during the invention of C. I blew my top.

I have since been informed by email (in blistering detail) that Mr. Firth is
not a random and seen by dmr's followup that he was correct. I retract my
insulting statements about Mr. Firth and accept full responsibility for my
error. Let the record at least show that I was as quick to apologize as I had
been to flame.

Mr. Firth was not among those who filled my mailbox. I shall try to take his
restraint in the face of provocation as a model for my own future behavior.

We're still waiting, Eric.


Jim on 11.29.05 @ 01:30 AM PST [link]


Monday, November 28th

HTML O' the Day



But there are other, more intriguing theories.
Whoop! There it is.


Penguins are not people, despite their natty appearance and upright ambulation.


I'm afraid this isn't going to work out. [Possibly offensive: AL]


When they left the hotel room Sunday morning, they left behind scattered Mountain Dew cans and Cinnamon Toast Crunch all over the floor, said Teyonna Upshaw, the housekeeper who cleaned the room.


"That motivated me to turn around and start running away," he said.


And when you stop crying and shaking at the sheer horror of that thought, I'll be here waiting. [Possibly offensive: AC]


Denial, resentment, bargaining, depression, blogging.


Click-to-Call


The KU faculty has had enough.


Creative use of drill bits.
Jim on 11.28.05 @ 07:09 AM PST [link]


Sunday, November 27th

20 years of Free Software


So I went to look, and just about 18 years ago, I submitted my first GPL-licensed GNU Emacs patch, a version of unexec for the Convex machine(s).


/* modified for C-1 arch by jthomp@convex 871103 */
/* Corrected to support convex SOFF object file formats and thread specific
* regions. streepy@convex 890302
*/

This, and some work Bob Miller and I did one night to reduce the number of system calls in the loop that reads characters was the end of Convex licensing Unipress Emacs. The GNU variant was so much better that merely substituting it into the ConvexOS release allowed Convex to close every open bug filed against Unipress Emacs.

I also helped Martin Streicher (who then worked across the hall, and is now the Editor-in-Chief of "Linux Magazine") fix the unexec.c used for Unipress Emacs.

After that I got busy on GCC. Quite literally I was working on GCC (for the Convex) in Dallas while Michael Tiemann was working on GCC for Sequent 32000 and G++ at MCC in Austin. My work on GCC ended up causing a lot of... trouble at Convex. Seems the manager of the (for sale) compiler product (and group) didn't like the fact that gcc had a better scalar code generator than the highly-prized Convex "vector" compiler, and that the marketing group would use GCC to 'win' customer benchmarks where necessary. Just after I left, the edict came down that you would be fired from Convex for using GCC in a customer benchmark, never mind that the proprietary compiler produced slower code in some benchmarks, and that any Convex customer could get a copy of the GNU C compiler. I'd also 'ported' gdb, the GNU debugger to the Convex architecture sometime during 1988. (I didn't sleep or date much back then.)

Note that this was about a year later, that Friday that was my "last day" was November 4, 1988. The Convex bits for emacs made the version 18.50 release. GCC for the Convex made the beta, or "compiler" tape in 1989.

I'd been using emacs for a couple years by then. I once managed to typeset the version 18 emacs manual on an Imagen laser printer (and hand it bound at the local Kinkos) back in October or November of 1986. (The manual printed as "Fifth Edition October 1986".) Several days later, I had the manual with me (I read it a lot), when I met Marvin Minsky when he was appearing with Carl Sagan at an "American Peace Test" anti-nuke rally in Las Vegas. They all went to get arrested at the Nevada Test Site the next day. I managed to get up to the stage after the speeches and had Dr. Minsky autograph the only thing I had on me, my GNU Emacs Manual. He inscribed it, "Best Wishes, Marvin Minksy, Friend of Stallman". I still have the book, as well as a "Sixth Edition" GNU Emacs manual, this time printed by the FSF, (with RMS riding the GNU on the front cover, and a "software hoarder" on the back) dated March 1987. It contains the "Clarified 20 March 1987" version of the "GNU Emacs General Public License".

ESR claims, "I was one of the original GNU contributors back in 1982-83, and I've been at it ever since." I find this curious, since RMS didn't announce "Project GNU" until September 1983, and work didn't start until January 5, 1984, when RMS quit his job at MIT. Work on GNU Emacs didn't start until September of 1984.

The first widely-distributed version (indeed this was the first public release) of GNU Emacs, version 15.34, was made on March 20, 1985.)

GNU Emacs 16.56 followed on July 15, 1985. The major change here is that all of the code from Gosling emacs was expunged due to copyright concerns. On Sepember 19, 1985, GNU Emacs 16.60 was released, and this release contained first patches from the net, including preliminary SYSV support.

What really grinds on me is that nobody has challenged esr's claim of having "contributed" "in 1982-1983" to the GNU project when work on GNU didn't start until 1984, and the "GNU Manifesto" wasn't relesed until 1985.

I can't find any elisp by ESR prior to 1991. (Now late 1988, see update below.)

Point in fact, during 1982-1983 rms was very busy single-hadedly duplicating the efforts of the entire Symbolics (nee: slimebolics) team in order to both keep Greenblatt's LMI apace and proprietary code off the MIT lisp machines. This super-human effort was the catalyst for what became "Project GNU".

Are we to believe that Eric Raymond founded "Project GNU"?

Enquiring minds want to know.

UPDATE:
Russ Nelson has provided a link to some evidence that ESR contributed to emacs as early as late 1988.

As I responded to Russ, I find '1988' (or even 1987) far easier to accept as an epoch for Eric's involvement with project GNU.

LATER:
rms wrote with:
"Eric contributed to Emacs development for a few years, around 1988 to 1991. I don't recall that he contributed substantially before 1988, but I could have forgotten something."

STILL LATER: I found where I'd finished a Convex port of gdb, by March 20, 1988, too. See also here.
Jim on 11.27.05 @ 12:57 PM PST [link]


further words on the crisis at Wayport


Someone (who wishes to remain anonymous) wrote in with some clarification on the issues at the Sonesta:
---
The background on the Sonesta issue...

I was at Wayport at the time that they "decided to stop servicing that location."

Wayports "decision to stop servicing" was aided in no small part by the sonesta physically tearing out every last piece of Wayport equipment and throwing it into a pile behind the building after reaching the last straw due to shitty tech-support that had been outsourced to a bunch of monkeys in Salt Lake City. Company by the name of Sento. These people were something else. I am pretty sure I even have a phone call or two recorded at home showing their brilliance at fixing an issue.

Anyway, the support became so horrendous, the Austin callcenter finally got a call and it was the manager of the Sonesta basically saying "We are finished with Wayport. If you want your shit, it is piled up out back. Otherwise, go to hell."

Solid business model. ;-)

They almost lost the Four Seasons contract for the EXACT same reason. Took some heavy duty meetings to rescue that contract even.

But who needs to pay attention to business travellers? Apparently the real money in wi-fi is at mcdonals. LOL

---
Now, a couple of things:

First, I really appreciate this kind of "insider" information from inside the company, even if you don't work there any more. Vucina is supposed to send all shareholders an annual "letter", reviewing the status of the company, but either I or Brett have had to remind him on several occasions that he's tardy. This further re-enforces the creepy "feeling" that I'd heard from someone (who I won't name) that Wayport had been "kicked out" of the Sonesta and nearly lost the Four Seasons account. Oddly, the Royal Sonesta still claims Wayport is providing service.

It also explains the discrepency between the two "halves" of the list of Wayport-served locations (though it says something even louder about whomever is writing the applications that extract this data.) Get to work, Keeler.

Third, the call center was always one of Wayport's "core competencies" at Wayport, at least while I was there. (I ran it at various times, as did Bill Hallett, who works for us at Netgate.) Rule one is that you don't outsource the things that make you unique, the places where you can "touch" the customer. At Wayport, the *only* time we dealt with the end-customer was when they had trouble, and we went out of our way to ensure that we could diagnose and fix the problem in a timely and curteous manner.

Fourth, Phil Windley is on the board at Sento. I know of, and respect Dr. Windley. Doc knows him real well. Interestingly, Phil mentions Wayport and Sento within a week of each other back in April 2004 on his blog, but doesn't link the two companies. This may be due to his presence on the Sento board (Sento is a public company), and Phil is just being careful. It may also indicate that the relationship between the two companies was over by April of 2004, or at least badly damaged.

Finally, I can't tell if Sento is still the outsourced call center for Wayport or not. I find a couple references to people who claim to work for Sento doing tech support for Wayport, but the pages seem stale.

Sento (or their PR agency) used to claim Wayport as a referenceable customer, but Wayport doesn't currently appear on Senao's list of clients.

In Japan, a sento is a traditional public bath house. A place for taking a bath. Someone took a bath, thats for sure.
Jim on 11.27.05 @ 09:05 AM PST [link]


Wayport: the continuing crisis


Doc continues the "pay net = pay toilet" metaphor with a damning critique of the (lack of) service available at the Royal Sonesta in Cambridge, MA:
Imagine being forced to unfurl promotional messages on a roll of toilet paper just to unlock the flush lever in a pay toilet that barely flushed in any case. At that same hotel, when I asked for improvements to the lousy bandwidth I was already paying for, the person behind the counter called over a manager who said, "What are you trying to do, get some email?" Wrong question. Especially at a hotel next door to MIT. Don't these upscale hotels have any idea how much that kind of stupid service pisses off potentially good repeat business?


What delights me about this is that the Royal Sonesta in Cambridge, MA was one of the original Wayport locations, though I can't find it claimed in Wayport's Master List (it does appear in Wayport's partnership with Know Where. Perhaps some massive database corruption exists inside Wayport's walls.

This hotel in Boston was also the first place where Brett's model started to fall apart. Seems the MIT hackers students were more than able to furnish their own NAT boxes for a demo in one of the meeting rooms, so they only had "one computer" connected to the Ethernet port in the walls. Brett was plenty paniced, even though the exact same code was being used on the linux box that controlled the IP interface to the hotel.

I'll just say it again, What will you do when its all free?"

After re-reading the PR from 1999, I recall that I was always going round after round with Brett about little pull quotes like this:

"Wayport is very happy to be introducing its services to the Royal Sonesta Hotel Boston," said Brett Stewart, chief executive officer, Wayport. "Located near the birthplace of the Internet, the Royal Sonesta serves guests and meeting attendees who value high-speed connectivity when they are on the move."

Since anyone who has a clue understands that the "birthplace of the Internet" either doesn't exist, or is UCLA.
Jim on 11.27.05 @ 01:52 AM PST [link]


Wednesday, November 16th

Scatalogic Professionalism


Scott Adams has a blog, wherein he recently posted on the issues surrounding body functions in a professional setting:

One of the AV technicians came rushing in and was startled to find anyone sitting there. He recognized me from the earlier sound check. Now he had a dilemma. I could tell from the look on his face that this visit wasn’t going to be the brief kind. The dressing room is tiny and concrete, with little privacy from the business area of the bathroom. I could practically see the gears going in his head.

Do I just act nonchalant and go pinch out a steamer in this tiny airless dressing room? How professional would it be to leave the keynote speaker marinating?

He hesitated. He said hi. He continued to the bathroom, then paused, put his head down and shuffled out as fast as he could, as if maybe he had only gone in to use the mirror to check his beard for anomalies.


This reminds me of the time I was in a Wayport board meeting, with the speaker phone going and the new investor (I won't mention his name, but he's a General Partner at a name-brand VC firm), who was, shall we say, "remote", clearly got up off the toilet (while talking) and flushed.

At least speakerphones don't carry video or stench. Yet.

Jim on 11.16.05 @ 07:31 PM PST [link]


A possible software revolution for free space optics


This could provide a real revolution for FSO.

Now, Pat Hanrahan and his team at Stanford University have figured out how to adjust the light rays after they have reached the camera. They inserted a sheet of 90,000 lenses, each just 125 micrometres across, between the camera's main lens and the image sensor. The angle of the light rays that strike each microlens is recorded, as well as the amount of light arriving along each ray.

Software can then be used to adjust these values for each microlens to reconstruct what the image would have looked like if it had been properly focused. That also means any part of the image can be refocused - not just the main subject.

Tracing the rays like this removes the conventional trade-off between the aperture size, which controls the amount of light that the camera takes in, and the depth of field. If light is low, a larger aperture will let enough light into the camera to form a clear image, but the laws of optics mean that a narrower slice of the world in front of the camera will appear in focus.


The major problem with most Free Space Optical systems is that the beam spreads; the more distant the connection, the wider it gets. Normally the way to handle that or to handle movement and alignment issues is to let the beam spread and take a loss in power, which usually limits the distance that the buildings can be separated. By using something like the lens array, you could do a sort of "beam-forming" for FSO links, increasing the effective distance, even for very low-power lasers, or even the LEDs found in commercial Gigabit Ethernet switches.

By allowing a larger aperture, the system can have more receive gain (more light collected at the receiver) and focused on the detector. This allows for a lower power transmitter to be used, reducing system cost.

Jim on 11.16.05 @ 07:13 PM PST [link]


Your future linux box


10Gbps Ethernet is making good progress lately. Fulcrum has a 24-port switch chip and both Vativ and KeyEye have Cat6 PHYs. Fulcrum claims 200ns latency and 1.2Tbps internal throughput, so essentially its switch-fabric is non-blocking except in the most constrained of cirucumstances.

10Gbps Ethernet over copper is currently limited to 15 meters (50 feet).

Chelsio's T210-CX 10Gbps Ethernet NIC is $795, and runs under Linux. They have been independently verified to deliver 7.9 Gbps throughput while using only 50% of a single server CPU... and Chelsio will continue to halve prices every 12 months .... So in three years, you'll find several sub-$100 10Gbps Ethernet NICs on the market.

Meanwhile, AMD has announced Quad core Opteron on its roadmap for delivery in 2007.

What does the "world's fastest linux box" look like in 2007? Likely a 32-way Opteron connected to 23 others via 10Gbps Ethernet, all running some whacked version of Xen and a single linux system image.
Thats 768 CPUs in a half-rack, for those of you counting at home. If each of these machines has but a pair of 1TB drives (lets assume that drive density only doubles in the next two years) then you've got 48 TB of on-line storage.

Assuming that these machines cost $10K each, and that the switch is in the noise, the entire rack will set you back a cool $250,000. While out of reach for most households, this does fit into the budget of even a moderately-sized business, especially one that needs some serious CPU power. Moreover, it will make for an entirely new class of 'hosted linux' businesses. Using software technology that is already available today, (and will most certainly be mature in less than two years) a near-future "service provider" could sell you (and two thousand of your closest friends) access to this "server farm" for a mere $20 per month. Thats $40,000 of income, balanced against a less than $8,000 per month payment on the hardware, and perhaps another $3,000 per month of co-location and base bandwidth charges. (You'll pay for any overages, of course.)

This leaves a close-to 50% margin of $29,000 per month to pay the small number people required to administrate such a machine.

Now, if we could find a way to get GigE bandwidth out of our homes there would be a lot more demand pull for these kinds of mini-clusters.


Jim on 11.16.05 @ 05:19 AM PST [link]


Tuesday, November 15th

HTML O' the Day


"We are hoping to have a small battery powered version on sale in the shops in time for Xmas."

I live in New York state, and I've tried searching through the weapons laws and could not find anything specifically mentioning a flame thrower, but still, I wouldn't want the cops showing up when you're playing with this.

Really? Who would have guessed?

The military people reached this conclusion, and they wrote it down on a memo, and then they classified the memo and Adel went from the hearing room back to his prison cell.

Alaska's whaling commission began researching new weaponry when the 66-member International Whaling Commission mandated two decades ago that more humane methods be developed.

... the world's ninth largest country yet still little known to many in the West.

Archimedes Palimpsest Project Summary

When your roommate has a death wish.

Google updates page ranking algorithms.

In pioneer days they used oxen for heavy pulling, and when one ox couldn't budge a log, they didn't try to grow a larger ox. We shouldn't be trying for bigger computers, but for more systems of computers. (Rear Admiral Grace Hopper)

"What would you rather have to plow a field two strong oxen or 1,024 chickens?" (Seymour Cray)
Jim on 11.15.05 @ 12:06 PM PST [link]


OSI, ESR, Halloween and Microsoft


In Christian countries, the Cathedral was the source of ecclesiastical power, and to some extent it remains so today. As a political power, however, its role has gone from all but central to mostly non-existent. Nobody is building Cathedrals anymore.

In Islamic countries, the bazaar, (which is seldom, if ever, co-located with any related job-producing construction projects), remains the center of ecclesiastical power. To sell in the bazaar, you need a license from a Mullah. In Islamic Republics, the bazaar is therefore the center of economic power, period.

Its entirely probable that the religious overtones in his choice of the term Cathedral was Eric Raymond's allusion to what I and other see as his pointage to the outcry from RMS (Playing the Pope), versus the more freewheeling style of Linux development. We've already seen that linux has had to become much more disiplined about acquiring copyright assignments for contributions as a result of the SCO mess. If you want evidence that both the bazaar style and the term "open source" are easily corrupted and/or prone to self-destructive, dogma-inspired holy wars, you need look no further than Microsoft and its most recent moves.

Yes, Microsoft.

Back in mid-October, one of OSI's boardmembers, posted this to her blog:

We voted unanimously as a board in April 2005 to move historical materials such as Eric Raymond's "the Halloween Documents" and Michael Tiemann's "A Case for Open Source" to the authors' sites for maintenance and to re-focus our efforts on increasing professionalism and credibility for both OSI and for open source worldwide. Re-focusing our efforts towards a positive future has been very successful. This year we've been invited to participate in South America, Asia and Europe at conferences and in private consultations with governments and local businesses that are contemplating open source licensing or policy issues. The OSI Approved brand is increasingly recognized outside the US. This puts OSI in a much better position to serve a growing constituency.

Yet the documents didn't move until October 17, 2005

And when did Microsoft go public with their new shared source inititative, which includes the two licenses that may be "OSI compliant"? October 18, 2005.

And only after did Microsoft meet with "a quorum" of the OSI board.

In fact, Danese has never mentioned ESR's removal from OSI's board.

Consider this little bit of history from Microsoft's own mouth:

Q: Why didn't you submit the IronPython license to the OSI for approval?
A: The OSI has provided a much needed aggregation point for open source licenses but there are elements to the OSI website that are strongly anti-Microsoft. They have every right to hold their own opinions and to post what they will on their website. For the time being, Microsoft has chosen not to work with the OSI. As we continue to expand our collaborative development efforts we welcome any and all constructive conversations with the OSI and others about how the various elements of the software industry can work together more effectively. For example, we think that establishing some consistency among source licenses is a worthwhile objective. With >55 "open" licenses and counting, it is clear that property holders have strong opinions about how they would like to share their property. All individuals and organizations must be able to reserve the right to choose the source license (or write a new one) that will work best for them. Yet, there would be significant benefit from some consensus on a few archetypical licenses that were well understood by all parties.

"... but there are elements to the OSI website that are strongly anti-Microsoft." Hmm. Posted April 5, 2005. Hmm.

Lets continue:

Software licensed under the PSF, BSD, PHP or Apache licenses for example can be used in conjunction with the IronPython license. Software licensed under the IronPython license may not be combined with GPL code (in this regard both IronPython.


Yet Tim O. the book god (no friend of the GPL or FSF) thinks that fine.

Perhaps this is the source of ESR's "We don't need the GPL anymore".

Eric is a racist dumb-ass, now publicly ridiculed. The OSI board have either sold-out or are trapped by their approach of defining a set of "license standards" and then being forced by their own lawyers to apply these licenses to all comers. O'Reilly stands to sell more books, so they're all in-favor of the move. Meanwhile ESR doth protest too much.

This move by Microsoft is just more "embrace and extend", and does not represent a softening of Microsoft's postion. Microsoft is poised to move to market its way into 0wning the very term "Open Source", and will then use it to position against Free Software and the GPL. Eric and the rest of OSI (along with Linux Journal The Leading UNIX, Linux and Open Source Authority Since 1983 and a whole host of other poseurs) have all but handed the fight to Microsoft.

Microsoft hates the GPL and will do everything within its power to destroy it. Only the FSF and the GPL are left to stand in their path.

Jim on 11.15.05 @ 06:44 AM PST [link]


Saturday, November 12th

Joy's Law


So Doc quoted me, but I think he missed the point I was trying to make.

I've seen Joy's law expressed in print lately as: "Most of the smartest people are never in your own company"- or ""No matter who you are, most of the smartest people work for someone else.", but the original I saw in 1990, when I was still a Sun employee was:

smart(employees) = log(employees)

Note that the base of the logarithm is unstated. Joy allowed that the base could be changed by the management style, culture or reward structure of the company. Still even if the base is quite small because everyone is getting rich and management gives daily hot oil massages to all the programmers, the results are not pretty. In this short table, the first column is the number of employees, the second is the log of the number of employees (base 10), the third is the natural logarithm of the number of employees, and the last column is uses a highly improbable base of 1.1

1.00 0.00 0.00 0.00
10.00 1.00 2.30 24.16
100.00 2.00 4.61 48.32
1000.00 3.00 6.91 72.48
10000.00 4.00 9.21 96.64
100000.00 5.00 11.51 120.79

Now, there are several interesting things happening. First, if you are the only employee, there are no smart people in your company. 'Nuff said about that. However at 10 employees, if your management team is epending all their efforts at keeping the smart people happy, there are 24 equivalent smart people. Likely your company is kicking ass and taking names. Note what happens as soon as your company has 100 employees though. Now less than half of them are smart, even under improbable conditions.

This is the diminishing returns I spoke of in the piece that Doc quoted.

Now, Google had 4,183 employees as of June 30, 2005. Microsoft has 61,000

4183.00 3.62 8.34 838.04
61000.00 4.79 11.02 1107.36

As you can see, Microsoft still has the potential for a few more smart folks, all other things being equal, though I find it likely that Microsoft's base is much larger than Google's. Its also interesting to note that the world's population is 6,446,131,400 and running the same test yeilds:

6446131400.00 9.81 22.59 2269.95

So Microsoft, which employs 1 in every 100,000 people of the total world's population may have about 1/2 of the smart people, and google has grabed a subtantail fraction of the rest, leaving the world with only 325 "smart" people who don't work for Microsoft or Google. This explains why the "law" doesn't hold outside of organizational boundries. Joy's law is really a criticism of the dynamics of corporations, especially those engaged in high-tech.

My point is that both Google and Microsoft make outrageous claims about having hired "only" smart people, but anyone who understands the dynamics of the hiring process knows that this can't happen as the company grows. Joy's law states this quite simply. For any company of more than a handfull of employees, hiring the next smart person get exponentially more difficult with each new hire.

Herein lies one of the secret strengths of the "open source" development process. Not all the smart people have to work for you, or even closeby. If the distribution of smart people is approximately uniform, it really is true that most of the world's "smart people" will never work for any one organization. The so-called "Open Source" development process understands and leverages this very fact. Anyone can contribute, and the contributions tend to stand on their own merit.

wnj, as those who know him refer to him, had another "Joy's law" once, it expressed that computing power of the fastest microprocessors, measured in MIPS, increases exponentially in time. The expression I saw was that the number of available VAX 11/780 MIPS (essentially a SPECmark of 1.0) was equal to 2^(year - 1984). Its late 2005 now, and so we should have 21 years of doubling, representing approximately 2,097,152 VAX MIPS equivalents.

The fastest SPECint2000 benchmark I find is 1771. Of course, the reference standard has changed, but I still don't know if a modern single-core Opteron is 2 million times as fast as the 11/780 I used in 1981.

I do know that an 11/780 would produce about 1700 Dhrystones, and this is what is often known as a Dhrystone MIP (one DMIP = 1700 Dhrystones), so that VAX 11/780 was good for one DMIP.

Recent postings of Opteron show that single core can produce 9191 DMIPS, or 15,624,700 Dhrystones.

If we were on-track with Joy's (CPU) Law, we should have much, much more.



Jim on 11.12.05 @ 03:34 AM PST [link]


Friday, November 11th

The wreck of the good ship Mabuhay


Today was the 30th anniversary of the wreck of the Edmund Fitzgerald.

I was 14 in 1976 when Gordon Lightfoot's ode to "a crew and good captain well seasoned" was released. It didn't get a lot of radio play, FM radio still being quite new to Las Vegas. I remember I liked the song because it was ran nearly seven minutes in length, and its tempo was great for slow dances with the girls.

The wind in the wires made a tattle-tale sound
And a wave broke over the railing
And every man knew, as the captain did too,
T'was the witch of November come stealin'.


Ah youth.

Years later, and another November storm. This time it was 2003 and the approaching shipwreck was happening in Spokane. The investors had decimated the management, and forced what remained of the company through two layoffs in quick succession. There was a new, unseasoned captain named Stalter, and he commanded through fear and outrage, not ability.

Our ship had hit bottom.

David (Lee) Roth, nominally a VP at this point, would wander the halls, holding his little political meetings, trying everything to preserve his own job and salary as long as possible. Whenever he would meet me he would start softly singing the lyrics from Lightfoot's song:

When suppertime came, the old cook came on deck sayin'.
Fellas, it's too rough to feed ya.
At Seven P.M. a main hatchway caved in, he said
Fellas, it's been good t'know ya
The captain wired in he had water comin' in
And the good ship and crew was in peril.
And later that night when his lights went outta sight
Came the wreck of the Edmund Fitzgerald.

This is true. Roth knew, as I did that the damage was fatal. His plan, unlike mine was to ride the ship as long as possible, standing on corpses of his fellow sailors as necessary to keep his own head above water.

Spokane doesn't have many tech jobs, and he had a family to feed.

Vivato appears to be about to release an enhanced version of its outdoor AP (not switch, AP). Vivato calls this the "VA2410 802.11b/g Outdoor Microcell". The FCCID grant date is 10/6/2005. Look at the internal pictures on the FCC site (FCC ID QLN-VA2410A).

I found this little conflict though.

The test report states: The EUT is an Outdoor Microcell 802.11 b/g transceiver. It contains two identical radios. Both can receive simultaneously, but only one can transmit at any time. The radio module is manufactured by Vivato, Inc.

Lets take the second statement first.

Vivato appears to have upgraded the cards in the former unit. The internal photos clearly show that they are using the Ubiquiti 2.4GHz card. Whats more, some of the photos are copies of photos used in Ubiquiti's FCC filings using prototype cards. (FCCID SWX-SR2)

I think the evidence is clear that the "radio module" is not "manufactured by Vivato, Inc."

Now lets take the "only one (radio) can transmit at a time" statement. First, and most damning, Vivato's own User Manual, filed with the FCC submittal, claims otherwise:

"Allows simultaneous 802.11b and 802.11g operation using two separately configurable radio interfaces." (page 18)

And, while the various documents assert that there are filters on board, these can't be channel filters, since: ""The VA2410 can communicat on any two channels in the IEEE channel set (although the default channel assignment of 1 and 11 should be used for best results). Both channels can operate at the maximum data rate of up to 54 Mbps." (page 17)

I also love how the install suppliment to the Users Manual stresses cross-polarization of the antennas.

Second, I can tell (but I know too much) just by looking at the pictures that the requisite logic used to control simultaneous transmission is not on the board. The CPU board (the one with the miniPCI sockets) is produced by Senao, and as already discussed, Ubiquiti supplies the radio cards. Vivato's contributions are the case, and the lightning isolation board. Even most of the software comes from Devicescape.

This from the company that two years ago had advanced and proven the radical idea of using phased array techniquies with Wi-Fi.

The only reason this unit may not transmit on both radios simultaneously is that the first radio that transmits quite likely sets CCA on the other radio. This is a pure artifact of the RF layer, no "engineering" was needed to obtain this result. However, as people inside Vivato well know, even this doesn't stop the real issue which happens when one radio is receiving and the other decides to transmit. Stopping this (and the beamformer used in the 11g "base station" prodcut), is a large part of Vivato's technical "art", but it is unused in this product.

This makes the second radio worse than useless.

This is Vivato's first new product since the outdoor 802.11g basestation was announced a year ago. When all you can produce in a year is this kind of trivial upgrade you're already found to be sitting too low in the water as the Three Sisters approach.

They might have split up or they might have capsized
They may have broke deep and took water
And all that remains is the faces and the names
Of the wives and the sons and the daughters.


Jim on 11.11.05 @ 06:49 AM PST [link]


Thursday, November 10th

Build your own Google


Here's something thats been filling my mind of late on the (too many) nights when I can't sleep.

Wouldn't it be possible to use a genetic algorithm to program a neural net such that the NN produced results largely indistinguishable from Google's "Page Rank" algorithm? If it is possible (and I think it is), couldn't Microsoft, or Yahoo! or some other well-funded source smash together a Google replacement (for search, anyway) and compete for AdSense revenue?

Note that such a competition would likely serve to raise, not reduce the AdSense rates. Websites would make more, while people who "buy" AdWords would likely pay less. Google (and its competitors) make a lower margin return, while the rest of us get rich in the process.

Yes, I'm discounting (to zero) all of that precious infrastructure that Google has built-up, along with their horde of ultra-smart people. Joy's law still applies, the number of smart people at Google is the log of the number of people at Google. Google may have a different base for the logarithm than Microsoft or the US Army, but they're still deep into diminishing returns with each new hire.


Jim on 11.10.05 @ 10:45 AM PST [link]


Intel + RF still equals zero


An ex co-worker of mine recently emailed:

Jim:

What is your take on the state of WiMax? What is it being used for and will it ever get any traction beyond backhaul? WiMax mobile? Anybody building chips?


I wrote back:

Intel has done a great job marketing/obfuscating between WiMax working at low transmitter power in unlicensed bands, where it can be no better than a mildly improved version of Wi-Fi, and working at high transmitter power in licensed bands. Can WiMax transmit much further than Wi-Fi in the cases where Craig McCaw has spent billions on licensed frequencies to broadcast at ten times the power? Certainly.

That apples-to-helium comparison is becoming more painfully obvious, and Intel's failure to protect the future of licensed Wi-Max (since Wi-Max has no future in unlicensed spectrum) from a vicious patent assult by Qualcomm will only accelerate the demise of WiMax, and quite possibly Intel.

Most importantly, free wireless hotzones are already being seen as a loss leader, even by Google. As Google leverages free Wi-Fi to push (at least) its local advertising business to critical mass, how likely are they to spend their time and money persuing licensed carriers, the FCC, and all the other baggage surrounding WiMax? Google owns fiber, they have no need for Intel's wonky broadband.

My prediction: WiMax will join HomeRF in Intel's collection of failed wireless efforts by mid-2008. Always remember, "Intel + RF = 0".

If you want to play at something *huge*, I have an idea for combining metro fiber networks with GigE bandwidth with cheap Free Space Optics devices. The idea is simple: fiber down the fence line (or inside an MDU) so everyone is connected at GigE speeds. Since the LECs won't let 'us' string fiber on their poles, and the munis want to be paid for same, we avoid the poles and cross the street with FSO.

Now we have a switched, VLAN-capable GigE network (architected entirely on Ethernet standards) connecting 100s to 1000s of households.

Turns out we don't want the network (broadcast domain) diameter to be > 7 for a variety of technical reasons, mostly having to do with time to convergence of the spanning tree protocol. 7 'hops' is 8 houses, or 8 blocks of condo/apartments. 8 office buildings.

Eight is number almost anyone can wrap their head around. Eight is enough. The Chinese think 8 is a (very) lucky number. Eight is "ba" in Mandrin, and "ba" also means "sudden fortune, prosperity". Confucianism has eight emblems, as does Buddhism. There are eight sides to the "ba-gua" (triagam) of the I Ching; eight "pillars of heaven", and so forth.

Next we need some bandwidth. Its not much fun to have this super-fast network with a soda straw connection outside of it.

Turns out we already have it. By leveraging the (very fast) network connecting all these homes, as well as the cable/dsl modem in many, if not each home, we can (in software) construct a device that allows us to leverage the total available broadband in the sum of all those cable/dsl modems.

People join the network so they can share files (TV? MP3s? Porn?) with their neighbors, but as an advantage, they also turn out to get much faster IP services, and possibly even some redundancy.

Remember as well that we can inter-connect these networks with FSO, so whole neighborhoods, indeed whole districts can be "wired" together with GigE speeds. Quite frankly, the applications don't exist today for much more than 10Mbps - 100Mbps anyway.

From there its straight-forward to invite Google and other metro fiber plays into the game, so the networks get some real bandwidth. Think 100Mbps for $20/mo, full-duplex to your home, with a chunk of (IPv6) address space (IPv6 already works on linux, freebsd, Windows 2K & XP and MacOS X.n.) You can host your own servers, your own PBX. There is an entire ecosystem ready to be built around real bandwidth into people's homes and standards-based IP/Ethernet networking. Google (or another metro fiber play) could inter-connect the districts (and eventually neighborhoods) with 10Gigabit fiber (Ethernet, again!) connections. Now you're sharing a Gigabit with seven of your neighbors.

Its built from the ground up.

Its entirely doable. We could transform the planet in < 5 years.

It takes some technology development, but nothing that is harder than "Sandhawk", which, I remind you, worked as promised, and cost as promised. Improving the Ronja (above) to GigE speeds at 100m is probably an early step.

Another thing: we can use an analog of the distributed CCA technique we developed for Sandhawk to build a distributed "WiFi switch" architecture that *preserves* capacity *and* security over a wide area. Yes, we can leverage 802.11n (MIMO) as well. We could, quite literally, make an entire city appear as though its was covered by the AP in your home and business. You could walk around and would not be able to tell that you've "roamed". Packets would hit your home gateway and be invited 'inside', so all of your home (or work) network(s) would appear to be local, even though you're miles away.

Does that blow your mind?

I've been considering writing up the "howto" for all of this and trying to get a meeting at Google. I have no desire to try to get it funded by myself. Its just too much effort, and I'm not very VC-friendly after getting handed Stalter as my reward for successfully swimming upstream in the HP-invested waters at Vivato for all that time. I did my part, and more at Vivato. As a company, we did may things wrong, and I did many things wrong, but we had a chance to deliver on most of the outrageous claims (sorry, I can't fix multi-channel until we can fix the client devices) that the HP-nauts wholesaled for eventual resale to the investors and customers.

For my efforts, I got anally raped at the hands of a 10mW intellect named "Don". Fuck the VCs.

Are you ready to get the band back together or was your query due to some WiMax startup looking at you for executive talent?

Jim

Jim on 11.10.05 @ 10:26 AM PST [link]


The good thing about being a guy is...


Brad,

It would be difficult for me to be any more miserable right now, I feel like the worst person ever. First, let me start by saying that I am truly truly sorry, and I hate myself for hurting you. Of all the people in the whole entire world, you were honestly the last person that I would ever want to wrong in any way. There is no excuse at all for anything that happened, so I won't even try other than to say all of us had WAY too much to drink, and I did a stupid thing. I can handle you being pissed at me, I absolutely deserve it, I can even handle the ugly words that were exchanged between us, what I can't handle is thinking that you see me as a different person. It is weird, I feel like I just went through a horrible break up or something. The world looked funny yesterday, I couldn't crack a smile if you paid me, there are songs I can't listen to, and I just ! feel beyond crushed. I don't know if you meant everything you said to me, and I am hoping that you didn't. I know that I was wrong on many levels, but I am also hoping that this is something that we can deal with. I know it sounds totally crazy and stupid, but you have come to play such a significant role in my life, I can't imagine my days without you. It is totally strange and weird to say that, and you could say that my behavior didn't reflect that, and you would be correct. I hate feeling like you hate me, and I hate feeling like all of your friends think I am a terrible person, because I am not. I know there is nothing I can say or do to take back what happened, but I just want you to know that fighting with you was just about the worst thing I could have ever imagined. It was right up there with one of the ugliest nights of my life, and I would give anything in the world to rewind and fix it. I am not sure if you will respond to this, part of me thinks that you won't. If not today, then maybe some other time. Also, thanks for getting my stuff together, although I think my sunglasses are still at your house, if you could keep your eyes peeled for them that would be great. I can't even focus or work today, I can't eat, I seriously feel like it was an ugly break up, and I am hoping against hopes that it was not that and you are not done with me.

Please don't cut me off, I really don't think I can handle that.

I am so sorry.

Elizabeth


Dear Elizabeth,

Thank you for your concern. I'll be sure to file it away under "L" for "Long-winded diatribes from drunken wh0res I couldn't care less about".

You did a stupid thing huh? No...doing long division and forgetting to carry the one is "a stupid thing"; Mixing in a red sock with a load of whites is "a stupid thing"; Bl0wing some guy in a bathroom for 45 minutes while I sit at the bar wondering if you're taking so long because you ate too much bran that morning isn't as much a "Stupid thing" as it is grounds for permanent removal from my social calendar.

To be honest, I'm not sure if it was more amusing that you went and degraded yourself in a public toilet not once but twice in a 2 hour span, or that you seemed to think that by saying "Well, I didn't F0ck him" somehow gave you a clean slate.

So forgive me if I couldn't care less if the world "looked funny" to you yesterday. Since your world revolves around blow dryers, golden retrievers, Prada Bags and Jelly Beans, I'm sure it must have been most unsettling to actually have to consider someone else's feelings for 24 hours straight. The good news for you is that my friends don't think you're a terrible person, they just think you're the average run of the mill cuum-guzzling blond who commands about as much respect as your average child p0rn collector. I could be wrong but, it's pretty hard to respect some B&T chick who comes out to spend the night at my place even though she's seeing someone else in New jersey and winds up tongue-bathing the ta1nt of anyone who decides 30 minutes of droning commentary on Colin Farrell's new haircut is worth putting up with for a hand j0b in the men's room. The good thing about being a guy is that when I eventually bump into the young lad who f1nger-blasted you on top of a towel dispenser last saturday, we'll have a shot and laugh our heads off about the time it happened.

By the way, for the amount of time you claim to spend in spin class you really must be doing something wrong to sport the thunder thighs you do. Watching you parade around my bedroom in a thong was a little like watching sea lions mate. Thought you might like to know.

PS. I BCC'd about 100 people on this email.

Talk to you never,
Brad

link

p.s. for those who aren't acquainted with the term "B & T" is Manhattan "in-crowd" speak for "outsider", as in Bridges and Tunnels, and those who use them to get to Manhattan. And yes, all the character substitutions were in the original.

Jim on 11.10.05 @ 10:11 AM PST [link]




your face here Home
Archives
Where I work
RSS 1.0 FEED
Powered by gm-rss History of sorts

What I (might) drive (soon!)

Greymatter Forums

Join FSF as an Associate Member!
November 2005
SMTWTFS
  12345
6789101112
13141516171819
20212223242526
27282930   

Valid XHTML 1.0!