Alan Hargreaves' Weblog

Sydney Open Solaris User Group

When Simon recently visited he discussed the formation of this group with a number of folk.

As a direct result we now have a (pretty basic admittedly) website and a sosug google group.

I'd encourage folk with an interest to sign up.

We haven't a lot up there yet, but I am confident that discussions will start soon.

Thanks to the xolinc guys for running with this.

I should probably also acknowledge Peter Lees too :)

OpenSolaris

Posted at 01:00PM Apr 27, 2005 by Alan Hargreaves in Open Source Solaris  | 

OpenSolaris is not Vapourware

There was a particularly insulting post to osnews that Fintan also took exception to.

I have to agree with Fintan's sentiments.

My feeling is that the crux of the insult is in the word "vapourware".

Wikipedia defines it thus:

Vaporware (or vapourware) is software or hardware which is announced by a developer well in advance of release, but which then fails to emerge, either with or without a protracted development cycle. The term implies deception, or at least a negligent degree of optimism; that is, it implies that the announcer knows that product development is in too early a stage to support responsible statements about its completion date, feature set, or even feasibility.

There is a similarity between vaporware and a species of hoax; both involve promoting a product or event which cannot later be produced. There have been a number of hoaxes in technological fields, wherein the hoaxter promises that proof of his offering will be forthcoming -- eventually. Examples include Clonaid, the Raelian company which promised proof of human cloning; or any number of perpetual motion machine "inventors". The distinction may be that in vaporware, the proponent truly does intend to produce the advertised product, while in hoax, he knows the product does not exist or cannot be produced.

Now by calling OpenSolaris "vapourware", there are a number things which you are immediately saying (that I suspect you might not be meaning to - I'll give you the benefit of the doubt).

Let's address these points in order.

1. ... announced by a developer well in advance of release, but which then fails to emerge

   OK, we announced it and have been talking about it for a while, but now we are within a few months of release. I'd hardly call that "well in advance" at this point.

2. The term implies deception

   To my knowledge we have not been involved in any deception and to imply otherwise is simple lieing and terribly insulting.

3. it implies that the announcer knows that product development is in too early a stage to support responsible statements about its completion date, feature set, or even feasibility

   See above, it will be out this quarter. I have yet to see a statement by anyone involved with opensolaris that could be regarded as not a responsible statement.

4. There is a similarity between vaporware and a species of hoax; both involve promoting a product or event which cannot later be produced.

   It's certainly not a hoax. I and the engineers involved in this would certainly not lend my reputation (which I live by) to such a thing. Nor would the Sun engineers involved or the more than 110 people who are not Sun employees who are a part of the pilot. I'd also be surprised if Roy Fielding would risk his reputation similarly.

5. wherein the hoaxter promises that proof of his offering will be forthcoming -- eventually

   I'd hardly call "before the end of this quarter" eventually.

While I can see where the original poster was coming from (opensolaris is not generally available), I think they could have chosen a better word to express their views than "vapourware".

opensolaris certainly exists, you only have to speak to anyone involved in getting it out there. There are a lot of us out there who both do and do not work for Sun.

I have to comment on the last line of the post that started this.

If Sun really is paying attention... STOP TELLING US ABOUT ALL THE GREAT THINGS SUN IS ABOUT AND EFFING SHOW US ALREADY! I've waited about seven years for Sun to make good on its promises.

Ummm what do you think Solaris 10 was? We spoke an awful lot about the great things in that. Hey, guess what? Solaris 10 is out and available.

Technorati Tag: OpenSolaris

Posted at 01:04PM Apr 26, 2005 by Alan Hargreaves in Open Source Solaris  |  Comments[19]

Things that make you go hmmmmm ...

Andrew "Tridge" Tridgell mentioned the bitkeeper debacle today at LinuxConf Au. If one is to believe what has been reported at GrokLaw, The Register and ZDNet Australia, one can only go "Hmmmmmmm".

I suspect we have not seen the last of this in the press.

Technorati Tag: Linux

Posted at 06:34PM Apr 21, 2005 by Alan Hargreaves in General  | 

Alan's Cardinal Rules for Customer Con-Calls

As a part of my work with the VOSJEC and occasionally for other PTS work I've had to be involved as a "technical expert" on conference calls directly with customers.

I'd like to think that I've built a reasonable reputation with those customers that I have spoken with. A big part of this is that I have a number of cardinal rules that I do not break.

Please keep in mind that these rules were written with a slant towards the kind of work that I do, they might not work for you.

I was chatting with a colleague in Holland (Dimitri De Wild) and he suggested that this would actually make a good blog topic.

1. Keep it technical

   Your part in this call is purely technical. If you are on the call, it is likely that you will have a good technical audience. Stay out of the politics. Taking part in the politics is not going to move us towards a technical solution. The people who handle upset management are the Service Account Managers (SAMs). There are times when it makes more sense to simply drop off the call and let the SAM deal with it while you work the technical issue.

2. Don't try to "bullshit" the customer, you will get caught out

   This piece of Australian slang may need some explanation. Simply put, it means something like "don't try to feed the customer a pack of lies", or at it's simplest, "don't lie".

   Quite simply, if you lie, you will get caught out and then you not only make yourself look bad, you don't do good things for Sun either.

   This is basic integrity.

   Another side to this rule is, if you don't know, say so. I've found that I get a lot more respect for saying "I don't know, but I can find out", than trying to bluff my way through something.

3. Don't say anything that you can't or are not prepared to back up with a technical explanation

   This is another integrity rule. Simply put, if you are not prepared to offer a technical explanation for something that you are saying, be it a proposed solution, a description of the root-cause, anything like that, don't say it. If you're not prepared to stand by it yourself, why should you expect the customer to?

I guess that the big thing to remember is that when you are on the phone to a customer, you are representing Sun. Anything that you say will reflect on the company.

Something else that I have found incredibly useful on such calls is to have a "back-channel" to any other technical folk in the call (eg irc, Instant messaging, ...). This allows you to not "wash your dirty laundry" in front of the customer. You can sound out the other folks on an idea before you present it, you can also present concerns over something that someone has said. I'm not saying that it should be anything sneaky or underhanded, simply useful.

Posted at 08:15PM Apr 20, 2005 by Alan Hargreaves in General  |  Comments[1]

Drinks in Sydney with Simon Phipps

Earlier today, I noticed Simon Phipps posting from Australia and remembered that he was actually coming out. Digging back through my email I discovered that he had mailed me and a few others about this and mentioned that he would be in Sydney on the 15th (today). I emailed one of the people that that email was addressed to (Peter Lees, one of our Solaris Operating Environment Ambassadors) and asked if Simon was coming up to Gordon to say hi. His reply was "I'm just going to pick him up, would you like him to come out?"

Of course I did.

As a result, a few of us who are involved in Open Solaris not only got to meet one of our CAB members, but as our Social Club had also organised a Beer and Wine on the Balcony afternoon, we also got to share a drink with him.

It was great to actually meet Simon. He took one look at me and said "Yes, you're Alan" (matching up the blog photo with the reality). We chatted about various things both including various Open Source issues/happenings, as well as other work and non-work related stuff; I gave him a brief tour of our lab (which almost has one of just about everything we sell) and then we went out for drinks on the balcony.

All in all it was a pretty good afternoon, topping off a colleague's presentation of his 5 year with Sun certificate, which involved a lunch at the local pub.

One of the reasons that Simon is in Australia is that a number of non-Sun folk have expressed interested in setting up Open Solaris user groups and he was meeting with them. The interesting part for us is that apparantly we will get leaned on heavily for content. Hey, that shouldn't really be a problem.

The only really disappointing thing was that James had to leave just before Simon arrived. Oh well, maybe next time.

Simon is heading up to the area that I live on the weekend, so I hope we have decent weather, as the Central Coast of New South Wales really is a very beautiful place. He's then heading of to the national capital (Canberra) on Monday.

Technorati Tags: OpenSolaris

Posted at 06:56PM Apr 15, 2005 by Alan Hargreaves in Open Source Solaris  |  Comments[2]

SCO Gives Sun Blessings to Open-Source Solaris

In September last year, SJVN quoted...

"Until we see what kind of licenses Sun is talking about and see what their plans really are, it's hard to say anything," SCO spokesman Blake Stowell said. "SCO is showing that it will fight for its Unix intellectual property rights with IBM, and I'm sure Sun sees that."

I guess that this article by the same journalist indicates that SCO got to see what we are doing and became reassured.

...during SCO's earning teleconference on Tuesday, CEO Darl McBride revealed that Sun had discussed with SCO its plans to open-source the Solaris operating system and that SCO has no problems with them.

"We have seen what Sun plans to do with OpenSolaris and we have no problem with it," McBride said. "What they're doing protects our Unix intellectual property rights."

There are some other interesting quotes in there.

Sun has slowly moved towards open-sourcing Solaris 10, or OpenSolaris, for over a year. To date, though, the only released components of OpenSolaris are programs, such as DTrace, which aren't parts of the operating system.

Actually, Dtrace is very much part of the operating system. If you download and have a look at it, you will see that a great proportion of it lives under usr/src/uts. Source code under here is what is used to build the kernel and the kernel modules.

Last fall, Sun officials maintained that the company had the appropriate intellectual property licensing rights to open-source Solaris.

Yes, we did, didn't we. Why is it the so many people won't take us at our word when we say things like that?

SJVN also mentions the Community Advisory Board announcement.

Anyway, that's one of the arguments that folks have been using against us doing the open sourcing that has been knocked on the head.

Unfortunately, things being as they are, and as James points out, we should now expect the hew and cry from the peanut gallery about how this proves that Sun is in cahoots with SCO and Microsoft to kill Linux.

Sigh.

I know that that there is an anti-Sun brigade out there who won't believe anyt hing that we say and will always believe that we have hidden motives based upon conspiricies. Unfortunately, there is little that we can do to change people who believe this kind of claptrap. About the best that we can do is to keep posting factually and let the more balanced folk make up their own minds.

Technorati Tag: OpenSolaris

Posted at 02:41PM Apr 14, 2005 by Alan Hargreaves in Open Source Solaris  |  Comments[6]

Andy Tucker on the CDDL

Andy Tucker, one of the folk involved in the development of the CDDL has refuted some of the misinformation that is being thrown about, especially by various journalists who appear to have other agendas.

He also talks about some of the reasoning for why some things appear in the way that they do.

A worthwhile read for anyone who has a desire to find out some facts about the CDDL.

Technorati Tag: OpenSolaris, OpenSource, License.

Posted at 11:36AM Apr 13, 2005 by Alan Hargreaves in Open Source Solaris  | 

Did Linus really mean to say that?

Various places are talking about the reasons that Linus is moving away from bitmover.

It appears that Linus asked Andrew Tridgell to stop reverse engineering bitmover's file format because BitMover was upset about what Andrew was doing. Linus paraphrases Larry McVoy as saying

You can compete with me, but you can't do so by riding on my coat-tails. Solve the problems on your own, and compete honestly. Don't compete by looking at my solution.

This appears to be a complete about face to me. It also appears rather hypocritical, and I am wondering if he really thought through the implications of what he was saying before he said it.

I could understand it if what Andrew was doing was to find out how bitmover does all it's clever stuff by reverse engineering the code, but from everything that I have been able to find, he was simply implementing an interface to the file format.

The Register likens what he told Andrew to do to be the same is if he told the Open Office folks to stop trying to work out Microsoft's Office file formats.

Closer to home, what does this statement hold for projects like NTFS support in linux (linux-2.6.11.7/fs/ntfs/ in Linus' tree)? Should we expect a statement about it's removal shortly? If not, why not? How is this different?

I'm not sure how anyone could see the cases as being different, unless you go along the lines of "but Bitmover were nice to us", which is an argument that should not hold any water.

I look forward to seeing how Linus clarifies his stand on proprietry file formats.

Technorati Tag: Linux

Posted at 12:26PM Apr 12, 2005 by Alan Hargreaves in General  | 

The Latest Brew

I noted back here about what I did with my latest brew, but didn't comment any further.

OK, It took some time to get the Specific Gravity down to something reasonable to bottle, but it finally got there. I cracked open the first bottle about two weeks ago.

Definitely interesting. It's certainly got a strong flavour. I find I like it best if I take a large mouthful at the beginning. I certainly like it.

My neighbour just let me try one of his latest that was very differeent. Along with the 1.7kg kit mix, he added the 1kg of sugar and 500g of honey. The taste is nothing short of amazing. Scarily drinkable. I think, I might throw some into my next brew :)

Technorati Tag: HomeBrew

Posted at 06:36PM Apr 11, 2005 by Alan Hargreaves in Brewing  |  Comments[1]

Pseudo-Terminal Allocation Before Solaris 8

My brother was having some problems with his Solaris 7 systems running out of pty's. We managed another workaround that enabled him to login and sort things out (which I might discuss later), but he'd been searching the web for information on this and turned up www.xegypt.net/articles/os/ptys.txt. I'm going to quote part of the page, but before I do, Let me say this right now, ...

DO NOT DO THIS. IT IS RISKY, WRONG AND YOU COULD PANIC YOUR SYSTEM.

The riskier way, and which I have personally confirmed to be working is done on-the-fly by writing symbols in the kernel image. Advantage? You don't have to reboot. As root, type this:

#adb -k -w /dev/ksyms /dev/mem
pt_cnt/W80
npty/W80 
$q 

The problem with doing this is that Solaris uses the boot-time values to allocate various structures associated with ptys, including a number of arrays (ptms_tty[] and pty_softc[]). If we change pt_cnt and/or npty on the fly, we risk working with memory beyond what was allocated to these arrays.

At best it will write over some data that we aren't using. At worst we could panic the box by writing to unmapped memory, or worse again, we could silently modify data associated with another part of the kernel.

The only way to increase the number of ptys on Solaris 7 and earlier involves a reboot and is the first method that he recommended.

The more reliable way to do this is to append the following two lines in /etc/system and rebooting:

set npty=128
set pt_cnt=128

The issue went away with Solaris 8, at that point pty allocation became dynamic.

Technorati Tag:Solaris

Posted at 12:24PM Apr 11, 2005 by Alan Hargreaves in Solaris  |  Comments[2]

Q and A with Simon Phipps about Open Solaris

Just noticed this article contining a Q & A session with Simon, specifically discussing Open Solaris.

One question and answer that jumped out at me was

Q: What do you envision the relationship between OpenSolaris and the Linux community will look like as it matures?

Phipps: I'd expect the two to develop a spirit of friendly competition. OpenSolaris starts on day one with innovations that are either absent from Linux or available in a much less evolved form, and I'm sure the Linux community won't take that laying down! Having real peer competition is good for both Linux and OpenSolaris and each will catalyze the innovation of the the other.

And this is exactly how it should be. Putting Solaris out into the Open Source Community adds to the community and the peer competition should act to raise the bar on Operating Systems in general.

All in all, this can only be a good thing.

Simon also talks about how he sees his role on the Community Advisory Board, his experiences on the pilot and thoughts about community.

Technorati Tag: OpenSolaris

Posted at 11:55AM Apr 06, 2005 by Alan Hargreaves in Open Source Solaris  | 

Solaris CAB - the Cat's out of the Bag

Today Sun announced the membership of the Open Solaris Community Advisory Board.

Congratualations to

• Rich Teer,
• Simon Phipps,
• Roy Fielding,
• Al Hopper, and
• Casper Dik.

A little more searching on my part showed that there is a list of bios accessible from www.opensolaris.org/cab along with a description of just what these folks will be doing.

Rich and Al were elected by the non-Sun members of the Open Solaris Pilot, Casper and Simon were selected for the two Sun positions, and Roy was selected as the independant participant from the Open Source Community.

Something that is important to note here is that of the five positions on the board, only two of them are held by Sun folk (Casper and Simon).

Ben Rockwood, Jim Grisanzio and Fintan Ryan also have something to say about it, as do Ashlee Vance at The Register, Stephen Shankland at Zdnet. and Elizabeth Montalbano at CRN

Yahoo has the press release.

Technorati Tag: OpenSolaris

Posted at 09:24AM Apr 05, 2005 by Alan Hargreaves in Open Source Solaris  |  Comments[1]