hpr-knowledge-base/hpr_transcripts/hpr1451.txt

Episode: 1451
Title: HPR1451: Jeremy Allison ~ the SAMBA project
Source: https://hub.hackerpublicradio.org/ccdn.php?filename=/eps/hpr1451/hpr1451.mp3
Transcribed: 2025-10-18 03:14:38

---

Hi, everybody. This is Ken from Hacker Public Radio.
In today's interview, I'm going to be talking to Jeremy Allison from the
Samba project recorded at FOSTEM 2014.
For some reason, I was unable to recover a audio track from my Zoom H2 for
this interview. So I'm forced to use the audio from the backup device,
which is why you'll hear some audio artifacts in this interview.
But as we say here at HPR,
any recording is better than no recording.
So I beg your apologies for the poor audio and apologies to Jeremy.
But I think you'll all agree that it is a very, very interesting interview.
So sit back, relax and enjoy the show.
And as I was walking past the good and summer of cold food, I heard a voice that I recognized.
And it was Jeremy Allison.
Jeremy, why should I have known you immediately when I walked past the booth?
I have absolutely no idea.
I don't even know about all of them, several projects.
You've been your host of the podcasting just going to say.
You've been a host of this week in tech on the Twitch network for a while.
But you've also developed one of the developers on the Samba project.
What else and many other equipment do you have available?
Well, sorry, you wanted me to talk about Samba, or?
Not just about you.
Oh, well, I've been doing, I've been doing free software for a very long time.
And I first started sending in hats to GCC, but the Sun 3865, if anyone remembers that.
My very first patch was sent to Richard Stormman and got rejected.
Very good.
Yes, it was great.
It was like, I have to say, it's like about onion, that wonderful onion headline,
crippled child gets answer from God.
No, it says God, you know, it's kind of like that.
You know, but I got an email back from God, Richard Stormman, you know.
Anyway, so yeah, I've been working on, I've been working on various open source projects.
How did you start off?
You were a kid somewhere in London, Jesse?
Well, I'm old.
Yeah, that makes two of them.
Might one of the kids, there weren't any computers.
Well, there were computers, but there was a sharp Zen something, it was Zen X80 base.
Zen X80 base was, I love Zen X80 base machine, and it was a sharp something 80.
That was the only computer in the school, there was one.
Yeah.
So I didn't really get into computing at all until my third or fourth year at university, really.
Where did you go to university?
I went to Sheffield, did physics and astronomy, and then I took a PhD doing geophysics,
which I was terribly bored with, and the guy who got me into the PhD,
what he really wanted was someone to fix his mass spectrometer control software.
And I was good at that.
I was terrible at the geophysics, I wasn't interested in this at all,
but I was good at fixing the mass spectrometer control software, which was based on PDP 11.
So had you any formal training and computer programming up until that point,
or was this on software?
I had an econ atom with 2K of RAM, which I took up to 12K of RAM,
and then I started writing myself things like a word processor and a little 3D graphics stuff.
The econ atom, if you remember, is this the English people I had?
No, not around the world, so.
Oh, so people, not...
Yeah, the bytes.
They may be a link to that in the short note, don't worry.
But people may remember, anyone who worked on the econ atom, it had
the most wonderful 6502 assembly language manual.
And so everyone at the time was learning basic, which I was, and then I ran into this section
in the middle of the manual, which was, it was like, oh, this is the weird stuff.
And I began to realise that this was what the computer actually used,
and so I got really, really interested in that.
And that weird stuff was essentially?
And that was 6502 assembly language, and so I started to learn assembly language,
and then did a side project while I was doing my PhD, I worked with one of the university technicians,
and he was, he was building control, mass spectrometer control boards, based on the 6800,
with the, with the assembler that we used to learn in our paper tape.
And so I wrote the control software for that, which, which actually taught me a great deal,
because when you're writing control software for a board, if you, if you make any mistakes,
whatsoever, the board hangs or, or just fails, and there's, there's nobody to debug it.
You just have to, you know, this is an independent board,
you just have to look at the code and work out where you screwed up, and so,
I think that's probably the only time I wrote perfect software, because at the end of it,
it just works, and there were, you know, we couldn't find any bugs in it.
It was very simple.
Yeah, that's what it did, yeah.
Yeah, but it was, it was, it was working software.
So that, that's how I started, and then after three years doing my PhD,
I ran away and joined the software house.
So, which one was that?
That, that was something called Kuma Computing, who wrote programs for the Atari ST.
Now, I had gotten, if you see my interview with Linus Torvals,
he and I both had a machine called the SyncLeg, well,
which was based on a 68.08 processor, 8-bit version of the 68,000.
I love the interview, actually, because he and I were stuck in South Palette.
Well, lots of stuff. We'd gone to South Palette Zoo to an Eskaper Linux event,
and I basically said, I will accompany you to the zoo only if you do me this interview,
and I knew I wanted to ask you about the SyncLeg QL, and it was great.
The YouTube comments, the comments on the YouTube were kind of like,
wow, I've never seen an interview with Linus that's so boring before.
I was like, it was like, screw you guys, you know?
I'm going to do an interview with Linus, and everyone asked him about how we changed
the world and all that shit. I'm going to ask him about the SyncLeg QL, because
that's what I like. So, I really enjoyed the interview, but I don't think anyone else did.
I do, actually.
So, I actually wrote, I wanted to come pilot on the SyncLeg QL,
so I got in the Atari, the SyncLeg QL operating system, written by Tony Tabby Kudos,
it's a beautiful, beautiful piece of work. I know we're going into old computer arcaneer here,
but it really was a nice piece of work, and the Atari ST ran something called STDOS,
which was a piece of shit. It was the port of MSDOS of the 68thound, and it was horrible,
but it was simply enough that the SyncLeg QL could actually emulate that OS.
So, what I did was I essentially wrote an Atari ST emulator, because they were like,
five or ten system calls, you know, open a file, read, write, change the directory,
this kind of thing. So, I actually wrote a SyncLeg QL emulator on top of QDOS
in assembly language, and then I could run the Metacompto C compiler, which was sold for the ST,
but not for the QL. So, at that point, I had a working compiler, and sorry, I needed a linker,
so I wrote myself a linker. This is a long time ago. And that's kind of how I got started,
you know, and in those days, the idea is, I mean, there was preparation software,
but nobody paid me attention to it. I mean, everyone was pirating games and cracking them,
and I moved onto the BBC Micro, which was a bit more robust, you know, and basically,
I got my start pirating games, I would like cracks for various games. And the funny thing was,
the only game I ever played was Elite, because most of the other games, I got more
enjoyment out of cracking the copyrights, not the fact that I played in the game.
Did you ever get into trouble for any of this? Oh god, no. And this is, you know,
it wasn't like I was doing a mass distribution network. There was no internet or anything.
This was something I was doing, kind of in my own home for the enjoyment of it.
Yeah. I had an awesome 75 slash 1200 board moment, so it would have taken a long time to upload
anything. What was that BT? I have some kind of bulletin board system, I can't remember the name
of now, but I was a member of that, and that was, that was, you know, everything was based around
things like CFACS, and that was a level of graphics that we had. Anyway, so that's kind of how I got
started. And then I moved to Kuma Computers now, I was running software on the Atari ST, and I
wrote the database called K-base, and this was, so it was a first time I'd implemented B-trees,
and the gem interface, and then came across the gem interface, which was really nice, I know.
If you compare that to early Windows, it was way superior, way way superior, really nice piece
of work. It had all these objects inside it. It was my first introduction to object oriented
programming, with, you know, with cobalt methods inside objects, and it was just a really nice
stall in C. Was that already in place back when the gem process? That was on the Atari ST,
and then they had a version for PC, but because the PC was on the 8086, or the 8088, they had those
bloody 64K segments, which were just, I mean, I was familiar with them because that's how the
PDP worked, and in fact the bug I fixed on the original mass spectrometer was the fact that
an interrupt handler had been placed by the linker inside a segment that was swapped out when
the interrupt fired. Which didn't help. So actually my main fix for that mass spectrometer control
software was telling the linker to move the interrupt handler into the global segment that was
always loaded. So yeah, there was a version of gem that ran on top of the IDNPC, and it was
shit. I mean, I don't know whether you remember, see it at that time. You would have things like
char, fast, char, star, far pointers, you know. So there were near and far pointers whether they
pointed within segment or not. And having been used to the 68,000, I thought, I just can't be
asked if this is the stupid thing I've ever run across. So I decided to stay on the Atari ST,
and let someone else who's actually a way superior programmer to me, put my program to the PC,
because I just didn't want to deal with the segmentation. And so I kind of stayed on the 68,000,
which was a flat memory space. Up until I moved to a company that was writing software on
sun machines, the early sun, I think it's on twos or threes at the time. And again, that was a 68,
68,000, 68,000 with a memory management chip. Anyway, yeah, and so I also run into the Amiga,
you probably have a bunch of Amiga fans on your show. Those are Amiga guys, they're weird,
God damn it. I always came to blows with this Amiga guys. I mean, you know, you criticize there,
it's like you're calling their child ugly, right? So why do you hit the Amiga? It was shit. I mean,
it was this tripod-based operating system. And it was just badly written, unusable crap.
So we had to port the CUME application suite to the Amiga. And so there was one real sort of
hardcore Amiga guy who was saying, oh, I'm going to, you know, let's port the ugly things properly.
And I looked at it, and I looked at the ST, and I thought, you know what? Why don't we just write
rewrite GEM on top of the Amiga interface? And then we can port everything as is.
So we had a race. So he did not support the applications to native. And I just rewrote GEM.
And I won by about six points. And we had a whole suite. So K-base was done, and then the K-word
guy and the K-spread guy said, well, we're going to use this GEM emulation because it's quick,
and we don't have to change anything. So we ended up, yeah, we ended up having,
did it run us faster? Sorry? Would it run us faster? Yeah, it was absolutely fine. I mean,
I mean, the only problems it didn't look native, it looked a little weird to Amiga eyes, you know,
but I hated that machine. This was, remember, everything was running on floppies,
and we saw we got our first hard drives to the Amiga. And you know, we had those, you know,
hinky, dinky doodle chips, or whatever they call it. You're going to get angry like this from
all these Amiga fags. Screw them, I don't care. So I had always so-called flashy graphics chips.
And I never forget running into a bug where if you dragged and dropped an icon, because, you know,
there's this whole sort of old dragon drop, wimpy interface stuff. So if you dragged and dropped an
icon, it was using the DMA engine to do the blitting of the graphics of the icon, which you were
dragging across the screen. If the disk drive, if the disk drive interrupt fired at the same time,
it would smear a copy of the blit of your icon all over the disk drive, because it was using the
switch. Oh my god. I know. We kept the losing disk drive. We kept from trashing hard disk drives
until we figured out, oh, well, you're doing that turn off the interrupt. Just, anyway, I hated that machine.
So, yes. Anyway, what are we very early days now? And then, how did you move on from that?
Oh, I went through various companies in England until I ended up at Sun. Yeah. Micro systems.
Now, Sun was interesting because it had the first, so I'd worked on Sun machines at Manchester
University in a bunch of other places. But I moved to Sun because you actually then had access
to the real Unix source. It's like, okay, this is the real thing. Because remember, remember,
kids, no open source back in those days, everything was closed up, everything was proprietary.
You wanted to learn how a Unix worked. You had to go and work for a Unix selling company.
It was the only way you could see the source. But this is after Richard Spellman sent the
email to HP about the printer? Yes. So, we were using GCC back in the humour computer days.
So, the GNU project had already been kicked off at that point. But there was no, there was no
operating system. So, all the operating systems were proprietary. And the pick your boys,
you know, and my boys was with Sun up. And so, you know, I worked at Sun and that's how I ended
up in the US. Sun brought me over. They shipped a version of what they called open windows.
So, we're moving to it next. And they had a new window system called open windows. And internally,
I'm sure they won't mind the being bought by all of them now, no one cares. Internally,
we used to call it broken windows three. It was so full of bugs. And I got temporarily
assigned to Sun's corporate technical escalations. Because working on the hotline, which is
what they call the hard-drinking, hard-party engineers who used to be at the front line.
And when customers called in, we were at the front line of taking questions.
So, that was the most technical job that you could do in the UK. Because there was no development
work done. So, to help broaden my experience, my manager said, okay, spend a couple of weeks
in the US and work for corporate technical escalations and work out how the bugs get fixed.
So, I went there and, you know, within a day of being in California, I thought,
I don't want to go home. It was mid-December. It was mid-December. You know, it was snow and
hail. I was Sheffield, man. You know, just the miserable weather. And then I'm out in California. I'm
driving up to 80 and it's 70 degrees this sun shining. You know, I see all these people on the
sun campus and having little Carol concerts. And I'm thinking, I don't want to go home.
So, I worked pretty hard to show. Look, I can fix these bugs.
What do they probably call it? Well, I think so much about them. So, what happened was
the TT manager said, oh, can we have him, please? My manager was great, because they said,
can we have him for six months? And my manager said, no, you can either have him if you're
high in permanently or, you know, or you can't have him at all. So, they did. They had me
permanently and I moved over. How did you find the move? Did you have any film to bring?
Yeah, I came over with my ex-wife. And that was rough because she couldn't work until I got a
green card. First few years were culture shock, because the US is different. The US is a different
place. What did you find the most difficult? I remember being shot that they kill people.
First time they were with an execution at San Quentin. I'm like, oh my god, it's someone,
just averages. You know, I still kind of have that opinion. I still don't think my death film
is a great idea. And I don't know the time. It's hard to, because I've been there now for about
20 years. It's hard to describe, but I remember being in, I was about to say parking lot,
because my, but now our car park, I remember being in a car park and looking around and all
of a sudden things felt foreign and there's no way to describe it. It was that the cars were
different, the, you know, the spaces were five times as spaces that you get in the UK. And it just
all of a sudden, for about 20 minutes, I felt completely out of place, completely foreign.
That has so much faded now. It's home. So anyway. Yeah, I've moved a few times, I don't know exactly
much about it. Yeah, if you haven't moved, you don't know what it's like, but all of a sudden,
especially if you move to a foreign country, there's this in events feeling of dislocation
that you get until you build a new home. Yeah. Okay. And you got your green card, obviously.
Yes, that was fun or an adventure, but let's not go into that. There's lots of shenanigans in
sun, which is a very interesting company. But yeah, we'll not go into those. Anyway, I got my green
card, ended up moving to a startup called Mantive, who was doing customer support software,
and I was there, I started out as their porting engineer and ended up being their network
architect. And that was really where, well, so that was where Samba started. And in case people
don't know what Samba is, because these days, everyone uses the web and nobody uses local
storage. But Samba is a file print and authentication server that allows Unix and Linux machines to
appear as storage nodes or printing nodes or whatever, or active directory now, on a local network.
So, you know, when you're watching your movies or music or whatever, on your myth TV,
oftentimes what it's serving off is a Samba server. You buy these little boxes from net
gear or whatever, and they've got Samba inside, and they serve your storage out to Windows boxes.
And so at the time I was working at Mantive, I mean, I have mentioned this many times before,
but I was working with PCNFS, which was a sun product, which was horrible. And for my sins,
one of the last jobs I had to do at sun was help maintain that thing. So I knew how bad it was.
And so I started to write something in C++, which is the only true language,
not that I believe that now, but at the time I started to write something in C++ and then
Trig and Trig or published the very first version of Samba, and it worked. And so I immediately ditched
my code and just started sending in patches. So, you know, so why were you supposed to be
producing something commercially to do the same thing? Well, no, I was actually supposed to be
porting Mantive software onto Windows, actually, Windows NT. And we did do, but the problem was all
the client, all the client code was on Windows, all the server code was on Unix, and it shared a lot
of code, and right now there was PCNFS, which had 8.3 filings, if you remember. And we needed
something that worked better to share the code, and so essentially I stole time from my employer,
like you did at the time, and probably still do people, to make Samba work. And every time I,
I got it to the point where we could replace PCNFS with it, and at that point we were really happy
because they stopped having to buy PCNFS licenses, they replaced all their PCNFS servers with Samba
running on SonofBoxes, and I saved the company a bunch of money. So they really care that I was
sending this code out because it wasn't anything that they were interested in. Yeah, and I knew the
GP, you know, I was familiar with the GPL, from the work I'd done at Son on the 3, 8, 6, I,
GCC work, and I'd read the GPL for the very first time when it was at Kumar back when we started
using GCC, and you know, coming from Sheffield, I read it, and I, oh, I get this, this is socialism.
Yes, come on. Wow, you know, I mean, it's not a dirty word, I know it's a dirty word in the US,
but it's not really a dirty word, you know. I just recognised, it was like, okay, this is what you
learn in these days, I say, kindergarten, but it's, you know, it's sharing, you know, you learn
to share, and that's what the GPL is. So, and when did you first move to church? God, I don't think
I met him till about five or six years after we'd started working together, because he's in
Australia, he was a PhD student, and I was in California, and I actually think the first time we met
was on the Microsoft campus. Okay, we're probably, so you were at Vantiv. Vantiv? Yeah, and
I'm from there, this is like your just audio resume, by the way. Yeah, that's what he's turning into.
So, what happened was I did a significant piece of engineering work at Vantiv. I added compression
and multi-thread support into Sun's RPC library. We were using Sun's RNC RPC as a remote
procedure call library to make Vantiv's product work, and so I put a, while my mum was actually
a member, I took a laptop over there and I was sort of coding by the side of the hospital bed,
and I got this working, and you know, it was a factor of 10 improvement in the network traffic
as in less, and so I back and presented it, and I remember them saying, well that's great,
but you know, we're already selling more of the product than we can handle right now, so we'll
keep this in reserve for when we need a big boost, you know, for our product cycle. Of course,
it never got released, and that code died, and at that point, I thought, you know what,
I've had it with proprietary software, and that actually, I think, is the last proprietary software
ever worked on. So, I went out at the time, and you know, I looked for another job, and the only
people who were doing open source software, well, free software at the time that they came open source
later, were sickness, my team in sickness, and so I went to work for them. They were in
Mountiveean, landings drive, he fought a walking distance from the Google building where that
worked, and Mozilla is probably occupying the same offices these days. I spent a lot of my life
up and down, sure, I'm Boulevard on Mountiveean off 101. Anyway, so, on the
Mountive building is now owned by Ignite, oh, one of the cloud vendors, who is a Samba user?
Anyway, so, yes, I went to sickness, and this is the funny thing, after I started working on it,
I thought what I was doing was I was, because I was very familiar with Windows,
or one of the few people who had a lot of experience with both Unic and Windows NT,
because I'd sidestep the lower versions of Windows until they came out with a Windows that
had no segments. This was the first version of Windows that was like Unic, right?
Oh, okay, I can work with this. And so, I was working at sickness on porting
Kerberus to Windows NT, this is back before the announcement of our HLH, which had native Kerberus,
and a little bit dirty secret at the time, is that sickness was actually experimenting,
me doing proprietary software. My Kerberus, this is going to be the first proprietary software
product, so I would like, oh, great. So, the way I worked on that was really interesting.
I started to look at porting MIT Kerberus to Windows NT, and sickness had this technology
that was written mostly by this, God, I have a program I called Steve Chamberlain, never
met the guy. Had he wrote some of the cleverest code I've ever seen in my life, just genius stuff.
Anyway, so he had this thing called SIGWIN32, and it was basically a
POSIX emulation library that's on top of Windows, and so I thought, well, I've done
Unic's to Windows ports, I know how to do that. Let's do something different. Let's
take Kerberus and try and compile it using SIGWIN32, and when it fails, when the compile of the
execution fails, let's fix SIGWIN32 and leave Kerberus alone. So, actually, what I ended up
spending most of my time at SIGWIN32 was doing SIGWIN32. Because you knew Kerberus was going to work
on it. Well, I mean, I knew that, basically, I knew that once we had gotten SIGWIN32 to the stage
where Kerberus just natively worked, we had a technology that was much more powerful than just
doing a native port of Kerberus. Yes. Because at that point, you can then take any other
Unic software that you need to run, you know, and remember, we're thinking about Windows as a server
operating system at the time. It's like, okay, then we can, like, shell-ing piece. Once we've got a
working SIGWIN32, we can take any working Unic server code and make it work. So, yeah, and so I
kind of took a vacation from Samba there for about a year or so. And how do you find SIGWIN
now as a project? So, I haven't worked on it for a very, very long time. It's a lot more sophisticated
and now it does a lot more. You know, if I have to use Windows and the dirty secret of the Samba
team is none of us really use this Windows much anymore. If I have to use Windows, putting
SIGWIN on it is one of the first things I do, because it gives you a working Unix environment.
But these days, most nobody needs to use Windows. Nobody needs Windows. I mean, you know,
if you want a nice desktop, just install Ubuntu. It works, you know, it's nice. It does whatever
Windows you think you might want it to do. Yeah. So, where was I? You've taken some time off
to do SIGWIN. Oh, yeah. So, then Microsoft announced Active Directory. It was going to be
Kerbera's base and the bottom dropped out of the Kerbera's market. So, at that point, SIGWIN
was like, well, what are you going to do? You know, what you could come work on GCC with all
our other contracts. Oh, God, I don't want to do that again. And at that point, a little
company called Whistle Communications in San Mateo said, how about you come work on Samba
for us, because we have this great product, and it leaves the plus will give you a 30 percent
pay rack. Because SIGWIN was cheap. A tradition that I think Red Hat has inherited to this day.
Anyway, I don't want to meet me in about Red Hat. I like Red Hat a lot. But SIGWIN was cheap.
And so, I... How did SIGWIN's involvement in Red Hat done? Red Hat bought SIGWIN.
Oh, okay, yeah. So, they picked up an incredible number of... Before Google, SIGNESS had the most
talented engineers I've ever worked with. I mean, it had just this... Steve Chamberlain left by
the time I got to SIGNESS. But we had Ian Lanz Taylor. We had Mark Horowitz. We had a bunch of
guys who had just genius-level coders. The guy who's doing... I forgot his name. Just so many of...
So many of the real names, a real power behind what's called Open Source. SIGNESS, you know,
nurtured and gave birth to those people. Really good. And you went to Whistle. Yes, went to Whistle.
And that was my first paying job to work on Samba. And that was great. You know, I mean,
that was when we first put the hot box stuff in, you know, opportunistic locking. Yeah, technical.
No, no. Well, essentially, it's a way of lying to clients saying, you have so
access to this file and guaranteeing that they can modify... Make modification locally and then
push back onto the server when someone else needs it. There's a bunch of complex work done
on that. And I actually wrote a column about that called The Startup Bicycle. I don't really
remember I read it. What are my favorites? When I talk about my time at Whistle. And so a start...
So it was a startup company. And a startup company is like a bicycle. And so the theory behind
it is, you know, start-ups of bicycles. But they're not regular bicycles. They're penny-farbing.
Okay. Yeah.
Engineering is the back wheel. The little one. Lots of the big wheel. The big wheel is marketing.
And most people don't understand that. I didn't understand that. I had the choice at the time of...
Well, I had the choice of joining NetApp, which I turned down, but they want to provide
But I also had the choice of joining Whistle or Cobalt. Now Cobalt, I looked at Cobalt and Cobalt
was a nice company, but they were basically selling a generic Linux on MIPS. They didn't have any
particularly special great technology. Whistle had everything. It had a great idea, a wonderful
concept. Apple UI engineers had the best of FreeBSD engineering team. Some of the cool FreeBSD
people. They even hired Kurt McEwsick to work on some of the soft updates that he did while I was at Whistle.
And Cobalt annihilated us. So what was the point of Whistle then? So Whistle had sold
this thing called the Interject, which was kind of... It was going to be like the HP laser jet or
whatever, but for the internet. So it was a little box that did file print authentication and
was your web server. And so you bought this thing via ISP. What a great idea that was,
because ISPs are such great salespeople, right? But it's through your ISP and it would be your
miniature web presence for a small company. And it was just a perfect idea. And the marketing
just was terrible. You know, the marketing person insisted we use Windows NT for our own website,
which of course all of the reviewers pointed out, and our competition. We've delighted.
But Cobalt had a cool blue plastic box. Do you remember how about...
I do, yeah. You know, they used to turn the lights on the dark and central off just to watch.
Just to see it, yeah. No, no, I completely understand. You know, the time house,
big plumbing mountain view. I remember going into the time house. And I knew we'd lost
when I saw that the time house had a Cobalt cube, which they were using, you know,
I don't know if there's something for their website or whatever. They had it mounted like a piece of art
where customers could see it when you walked in. And I'm like, oh yeah. They really get it.
And, you know, they had all these risqué adverts that they used to get criticized for. But the
marketing around that thing was just some bought them for something like $2 billion. And then
just promptly pissed it down the time and the tubes. I never heard anything about them after that.
I thought I would take them to the whole brand and do something with them. No, it was some.
But you have to have some bought them. That's what I mean. I expect some to continue the brand and,
you know, you never worked at Sunday, did you?
Ah, no, no. I mean, I would have loved to have been a Cobalt and taken the money.
But no, once some bought them, they were doomed. Just like everything else some bought.
Yeah, so, I don't want to become obvious that Cobalt were winning out a couple of business.
I don't know. I wasn't at whistle that long. Probably about six, seven months later.
It became obvious that we were just being annihilated in the market. One of these awards,
the award used to say whistle. The whistle in jazette is the nicest box you can't buy.
Because you have to buy it through your ISP. We wouldn't sell it directly. And of course,
no ISP had any clue how to sell anything.
So, in the end, I ran away and joined the STI.
Yep. An STI at the time was exploring, doing Windows machines.
What a wonderful idea that was.
With, you know, I haven't mentioned any names, have I?
Because I realize there's a bunch of stories I really shouldn't mention names.
Oh, thank God for that. Okay. Yes. So, SGI experiment was experimenting with doing Windows
clients and burned about $350 million on that little adventure, I think, which promptly pushed
the company down the tubes. Anyway, so I was there to do Samba on Rx in order to provide
the service support for the new Windows machines that SGI was selling.
And I think we did a decent job. It was the first 64-bit port of Samba.
It was the first time we moved some of the Samba support into the kernel. You know, it was a lot
of really good work that got done. That was SGI. How was the Samba project managed to
contribute to us? So, it's a boxed on the GPL-based project.
And so, we used to only accept individual copyright, which was quite interesting.
I don't remember when I moved to SGI. This would be before I got very careful about
assigning copyright, etc. Before people took this stuff seriously,
but I'm going to SGI and the standard Californian employment contract basically says,
we own every bowel movement, you know. And so, when they gave me this, I said,
oh, well, you're not going to own Samba. So, I sort of scribbled on an employment contract
and said, except for Samba and cost a few things out on hand, and then it back.
And they said, you can't do that. You know, we saw it with a much smaller company.
See, they never had any people with this kind of thing. So, I did that, and I sent it,
you know, I said, well, that's what I'm signing. So, I signed it and gave it back.
So, well, the lawyers were getting too busy about this. And I can't remember three years later
or whenever it was on left. No one would have definitely gone back to me.
One of the advantages of a large company where no one knows what anyone is doing, as I suppose.
So, we always had individual copyright. So, normally, all you would do to contribute to Samba
is just send a patch into the mailing list, and someone who has commit access would commit it for you.
These days, like most open source projects, we've actually gotten very
serious, and this is one of the things I learned at Google, because Google, as you know,
is a very good engineering team. And so, these days, to get things into Samba is not so easy.
We cannot now accept corporate contributions, and that's actually part of moving to GPL version 3.
All right. Let's go back to that. Okay.
Because there's a whole story. You were at CGI. How did you do SGI?
SGI, sorry. SGI.
And then, you know, I was working quite closely with a lot of the other Linux companies in the
valley. It was a very strong Linux. The Linux, kind of correct my name, the Linux users club
basically, Silicon Valley. They essentially became VA Linux. And so Chris,
Devon, who's now my boss, basically, he kind of said, leave it. Leave SGI. Come join us. Come join us.
As he did with Google, really. And so, in the end, you know, I'm only human now.
New VA Linux was going to go public, and I thought, okay, you know, why not? Yeah,
if I went and joined VA Linux. And I made no money whatsoever out of it. I wrote it all the way up,
and I wrote it all the way down. Yeah. Because I'm an idiot.
I think some people might know that the VL Linux was one of the first Linux companies to go
public. Yes, I think Red Hat was first. Red Hat was first, and then VA Linux was next.
VA Linux still, I think, broke the record for the single day biggest IPO game.
Yeah, you know. Oh, why doesn't he have so much stuff? Probably something.
So, why did it? Why did it tank? Well, eventually, why tank was because
what we were doing was we were selling generic boxes. And that really wasn't the value.
VA Linux, I think, had a great deal of values. And what tank was basically, this one, the internet,
but first internet bubble burst, you know, in the 2000s. And so all the customers went bankrupt
to enjoy it, all the people were buying, you know, thousands of thousands of servers. And the value
of VA Linux wasn't really the hard way. A story I heard was that VA tried to sell its hardware
to one of the large vendors. And at that level, you know, you get 15 minutes in front of the VP
to make your case as to why this is your value. And what they came in and basically said was,
we have these great boxes, this great technology. And the VP said, we already have that. Next,
what VA had was it had one of the best Linux engineering tickets. So we had Ted Cho,
we had Andrew Churchill, we had, we had all these, you know, what they should have gone in
and said was, yeah, we have generic boxes that you can replace, but we have your Linux
kernel team, you know, we have your Linux, where we had a network of storage products,
that network of storage products became readiness, which is now on by NetYear. Yeah, so I mean,
you know, there were ways to make money out of this thing. Anyway, so I can say, I wrote it all
the way up and down again. But VA was a guy, I'm still friends with so many people from VA,
that was such a great ride. Everyone should do that roller coaster ride, just once, you know,
just just just just to see it go up and come back down again. Yeah. And after VA Linux,
so I went to HP. HP wanted to do a printing of plans based on Samba. And so that was one of
the boatloader the printing work got done. Yeah. And yeah, I wasn't, you know, the years in the
wilderness really working on printing. Yeah. I would say once a printing card was for Samba,
or the Samba community. It was really cheap. You could run, so remember, HP makes money on ink.
Yes. More printing, more money. Yes. All the way to it.
Every little software. Yeah. He printing those flyers. So the idea, the NHP printer plans,
which was very, very popular, was you could buy a small rack mounted server. We would run on
a tiny box with 256 megs of memory. Yeah. You couldn't even boot Windows NT online,
which was the only other alternative. So we implemented the Windows printing protocols. And,
you know, you could, you could run 300 simultaneous print jobs through that thing. Yeah.
So the value add, and before we got cancelled in HP, what we tried to push to the management
is, we are the printing platform. You know, move to Linux for your print servers, put Samba on it.
Yeah. And you have a platform that you can sell other people to write apps. You, you know,
beef up the box a little bit, get people to write apps that run on it, let them work as filters
in the print process. And you have a platform, you know, just like anything else, it's like Android
or whatever, that other people can write apps on. And, you know, that way you will own,
as soon as it leads to Windows box, you will own the print process. And you can make it really
easier to print more pages, which is where the money is. And anyone in the park with Windows printing
at the time, it was torturous. Oh, it was horrible, horrible, completely undocumented too.
And in the end, they canned it and tried to replace it with a box running Windows.
You know, that would actually run embedded versions of Office. And the whole project was canned
after a year, because they couldn't make it work. Anyway. So, your, your whole resume seems to be
disaster. Yeah. Oh, it's a litany of disasters. And until Google,
everywhere I went, you could make money by shorting the stock everywhere. Jeremy's gone there,
short the stock. It's going to time. Yeah. So they're talking Google. Well, not ex-except for Google,
because I am, you know, I am such a minor car bunker on the side of Google, but, you know,
even I can't run them down. Did you stay on the stage for your, uh, so I didn't age
for a while, but after they cancel the project, I moved to Navelle. And so we did a lot of work,
Navelle at the time was pushing the Susie desktop. Yeah. So we did a lot of work to make the
Windows integration offline log-on and all that kind of stuff work. And that was a lot of fun.
And that was, that was when you got, uh, active directory integration. Yes. Well, we,
we hadn't before, but we were, we really made it work at that point. You could do offline log-on,
you could do, you could take, you could do with a Susie Linux box exactly what you could with a
Windows, um, laptop, which is still needed to control it. Yes, we still didn't have the
domain control. Um, so yeah, I mean, that was, spent a lot of time on that. And then, you know,
eventually I left after the deal with Microsoft, which, you know, as this, on the historical
record, I was not too chuffed with, really. Um, you got precise for that, but if you,
you, that you took too long to, you stayed at your level too long after, and then you moved to
Google and people were saying that it was, uh, you know, you should have moved earlier or something,
what do you say to yourself? Did I? Okay, I'm never going to say that before.
I'll get you. Oh, no, I mean, I, I stayed in the valley and argued against it. I mean, I,
I saw an original, uh, version of the agreement before it was public, and, you know, my response
to the lawyers was, tell me why this isn't a GPL violation. And the response I got by was,
please delete all copies of the agreement that you've got. Okay, very good. You know, uh, and so
one, once that happened, basically, I moved straight away. So what was the, what was so
grievous about the thing? Um, essentially, it was a way to try and get around, um, GPL V2,
by instead of calling it a patent license, they called it a covenant not to sue. Okay. So it was,
it was essentially re-raining the deck chairs on the, you know, um, so that it, it, it was called
something else and what it was. It was really a patent license agreement. And the whole
point of GPL was that, you know, that you, essentially, you, you can't do side patent agreements
that violate the GPL. And that's very explicit in GPL V3. How do you think that those worked out
from above? I don't, I mean, to be honest, I don't think it hurt them. I think again,
there's a lot of credibility with companies who found that sort of thing important. I don't
think it helped them that much. Um, you know, um, so, you know, I, I still have, there's still some
great sympathy members at Navelle. I still, I, I still, you know, Greg Krahartner's at, was at Navelle.
Maybe still this. And it was succumbed to the Linux foundation. But yeah, no, I, I, I get on with
Navelle 5. I don't have any, any, you know, long standing issues. And then it was just throw your CV
as Google on the hard drive. No, no, no, no, no, no, no, no, no, that's, that's not how Google
works. That's me. If you've ever been for a Google interview, you'll know, yeah, okay, so you know
that's not how Google works. No, I did a Google interview just like everybody else does. And yes,
that's not easy. I guess not easy. How have you found it so far working for Google?
I love it. I mean, it's, the best way to describe Google is, I once worked at Manchester
University. And, you know, I remember one of the questions that the guy, the one of the things
that the guy said to me when he was going to offer me the job, he said, you do realize you're going
to have to come into work every day. I was like, yes, yes, it's called working. I was expecting that.
And the guy said, because he said, the last guy on your job, he only came in three days a week.
And that's not going to be acceptable. So that's not to say that you only come in three days
and we get Google. But what I'm saying is how easy the environment was at Manchester University.
Well, work at Google, because I enjoy what I do, I mean, it's not that I don't have any pressure,
but I do, and so do a lot of engineers at Google, but it's hard to call it work.
Yeah, it doesn't hit for your hobby company.
Yeah, it's actually, it's fun. And that's, you know, you can't ask for a better job than being
paid to do what you probably would be doing even if you weren't paying it. So even though, yes,
you know, I put in the hours, you know, I won't say anyone at Google doesn't work hard,
but everybody works hard. But it reminds me of being at a university, it reminds me of being
at a campus, because people are there doing stuff that they love in an environment that is
designed to make it fun. Okay. And it's, it's, it's great. It's, yeah, it's the best you've
ever had, basically. Cool. One thing I think, which earned you a lot of respect, is the,
or Microsoft versus the EU thing, and getting the whole documentation for, for Microsoft products
released and hold lines for Microsoft. I think your team were involved in that whole,
can you tell us that story? Well, we were the only ones that couldn't get bought out.
Because we didn't own the code. So what was the whole deal?
Well, so some originally bought a complaint against Microsoft, because
long, long story, many, many years previously, I had, I had dinner with a son of VP trying to
say, look, you guys should throw in your lot with Sander, you should work with us, you know, and he,
he said, that's very interesting. And then he went away and he bought a product from AT&T
called Anavancel of Units, which was basically a port of the Windows code, and said, right,
well, that's going to be our Windows integration strategy. So as soon as son licensed that from AT&T,
Microsoft pulled the license from AT&T and said, we will not give you any further updates to the
AT&T directory stuff. And then AT&T sued them, because Anil, because AT&T asked me to be an
expert witness, and then Microsoft said to AT&T, he is $250 million go away, and leaving son in
the lot, so son complaint to the EU, and then a boatload of other competitors, Navelle was one,
whatever, everybody joined in, and so Sander was, you know, the Sander team represented by the
Free Software Foundation was another one of the complainers, and the end of the thing, I think
we were the only one left, because it was like, you know, son, oh, here's $2 billion go away,
Navelle, here's your agreement, you know, all of, you know, slowly but surely, all of the smart ones
took the money in ran. No, no, it's true, but what would you do if someone offered you a billion
dollars? You'd probably think about it, right? Anyway, so, is that all for money? No, we did it.
No, well, so my, the only thing I thought of it would be hilarious, if the offer is
funny, if we turn around and donate it to the Free Software Foundation, so having markers off
basically paid for the FSF, a billion dollars or something, you know, when people at the
software would be hilarious, but we never got offered money, because I think, I think they knew
that, you know, we couldn't accept it anyway, it's not. That reminded me of the story from the
early days when, when Tridge, when, you know, PCNFS vendors weren't doing so well, because Sander
was doing pretty well, and, and one of them turned to Tridge and offered him $40 million,
so can we buy Sander from you for $14 million? That was before they, people even knew what the
GPL was, you know. So, so we don't own the code, and this, this is one of the things I'm talking
about where we only accept it, personal copyright, so the code is owned by the individual engineers
who put it in there, and that's great, because it means we can't be corrupted, because we don't
earn it, you know. You own your bid plus. I own my little bit, and all the people own their little
bits, but, you know, I can't take the money and run, so the great thing is, you know, I can't
be tested in that way, so I could make fun of, you know, I would have taken the money and run,
because I'll never know, because I could never do that. Yeah. Why were you even trusted?
Well, essentially what was happening was that because of the close time together of the
Windows plant and the Windows server, and this was, I'm not going to say that Microsoft did this
deliberately. I don't think that they did. I think they considered a happy side effect of the way
they were doing their engineering, that it was very difficult for anyone else to re-implement what
they were doing. And the reason that it happens that way is because they were using a complex
suite of remote procedure call protocols, and they had a lot of legacy code to support. So, for
instance, you make a call and it gives back an error, then it falls back to doing something else,
and then it makes a call, which gives a different error, and you see what I mean, you can't just
implement the things on the wire. You have to know what the right error code is to make the
clients do a different thing, and that's not done maliciously, that's simply because when you run
the code that's been auto-generated from the RPC mechanism, that's what you get. And remember,
because they have both plant and server, and they didn't use to test against anyone else,
then what you end up with is products that are so tightly bound together that it's impossible to
substitute one or the other. And although there was some things where it was suggested that
you make it deliberately, you know, windows awesome finished until the sambal was broken.
I hurt those rumours, actually from some people, I trot, I don't think that was ever an official
policy, I don't think. I mean, you know, maybe as much as remember the windows isn't done until
what was developed, you know, the developers broken you. Yeah, the early dots. I really don't think
it was deliberate policy. I think it was a side effect of the complexity. But what it meant was
that, you know, we did a decent job, but we couldn't do a good enough job in order,
everyone else was being forced out of that market. So from from their point of view,
if a server was connected to the client or a client would continue to, it would try the
latest API and then fall back and fall back and fall back. So you have it's more an effect of
backup and policy than it delivers attempts to lockout sambal. Yes, I think that's the case.
Okay. And, you know, remember because you control both clients and server, it can be as complex
as you like. The dance of interoperability can be, can be a, we produced a paper in the EU case
that was called Microsoft's Web of Interconnected Protocol. When we tried to explain that it wasn't
enough just to get one piece right, you have to get every single piece right, including the error
mechanisms in order to make things work together. But you were here a long time, never to call that
reverse engineering. Network analysis. So reverse engineering is something that we never had to do.
Reverse engineering is when you take compile code and you decompile it and you look at the
assembly language, you know, you try and turn it back into C or C++. We never did that.
We looked at what happened on the wire. I mean, in fact, so the really funny thing is, when
after we were cooperating with Microsoft, the Microsoft and Sam were having a really good
relationship at the moment. So I don't want to, you know, I don't want to go too much on the
murky and unpleasant past because it's kind of murky and unpleasant. Yeah. And it's not,
it's not how things are these days. But one of the funny things that happened once we started
collaborating and cooperating together. And remember, Microsoft fixed bugs in Windows now to make
it work with Samber. They have fixed client bugs to make it work with Samber active directory
and the main control was. You can't get greater cooperation than that. But back in the days when
we were trying to, when we were first working with introoperability documents, what,
what Microsoft, what we found hard to communicate to them is, look, because they kept saying,
well, this API does this on Windows and we would say to them, you don't understand, we don't
care about your APIs. We don't even know what your APIs are because they describe everything in terms
of APIs that run on the client or the server. And we said, you're completely at the wrong level.
All we know is what's on the wire. How that's layered on the server is irrelevant to us. You know,
if we make a call and we see that error message, a certain error message coming back, we don't care
whether it came from the anti-FS layer or the anti-curnal layer or the SMB server layer. It's an
error code we see on the wire. We have to know where, how to respond. Yes. Because a lot of the
time we were asking questions like, why do we get this error message back? And the SMB server
guys would go, we don't know. We passed that amount to anti-FS and we just passed back what we get.
You know, and the API says this and we're like, well, we don't care about you. So when we know,
we now understand each other. That's really good. It's like, okay, no, Jake. I think a lot of the,
a lot of the issues, you know, after they were forced to release the documentation was that.
Well, so the interesting thing about the quote being forced to release documentation,
Microsoft made a conscious decision to do it right. They didn't just release the documentation
that they were quotes forced to by the EU. They released everything. They essentially
decided, I think, as a corporate decision, that anything going on the wire that they could use,
they would now document. So they did a 180 degree sea change. You know, in some ways,
in some way, I can just see them laughing because they essentially said, be careful what you ask for.
Here's three million pages.
Oh, it's like documentation that tell you exactly what we do and how we do it. Now go and implement
that. There's good bugs. You know, I mean, in some ways, that's kind of what happened.
So yeah, they did the right thing. They documented everything. They documented it as well as they
could. You know, they have teams in China that do various work to make sure that the documentation
is correct. So they have team in China that takes the documentation and then tries to
be used as a product, but does the same thing as Windows does with it. So yeah, they're doing the
right thing on that. Just why is Google hiring a sambal team? Or is that the top-seeking Google
project? I could tell you that, but then I have to kill you. So Google does use sambal
internally in various things. So, you know, I mean, it's useful to them, otherwise they wouldn't
hire me. That's kind of, you know, what you can say about that. So, and we were talking also about the,
but well, plus, plus I do, you know, unsource work, I mean, the open source program office,
you know, I do work like, I mean, I'm not here for sambal. I'm not giving a sambal
talk at this conference. You know, I'm here to help out with the summer of co-pay.
No, no, no, I'm Belgian beer, chocolates. Well, I'm not in the same depth to those things,
I'll admit that, but hey, you know, I'm, I'm staffing the booth. I'm talking to attendees, you know,
walk in the walk. I'm helping out, yes. So, is there, is there anything else in this
interview that I have not covered? I don't think so. Do you watch the last thing about sambal before?
Yeah, sambal before is an active directory in the main controller. This is amazingly cool.
It's only taken 12 years to do. I mean, one of the reasons of that is there's so many moving parts,
there's so many components. So, yeah, it came out last year. We are
rapidly developing it, we're implementing new, the new SMB3 protocol that Microsoft released
were also improving the active directory. The main control was to allow Windows clients to fit
into it better. It's, it's very cool. You should check it out. And is it stable at this point?
Yes. And it does everything in a standard Windows active.
It doesn't do multi domain, multi forest. I can't remember. I'm more of a file server guy,
so they doesn't do the multi domain replication, but that's something it's been worked on.
Okay. Is there anything that you would have loved to have done in your life changed differently?
Uh, joint robot maybe. No, I have a pretty good time. If I would, so if I could do one decision again
with the technology that we have now, I would not use C for Samba. I wouldn't use C++ either.
I would use Google Go. There are a bunch of talks from Google Go engineers.
Tomorrow, actually, I think it's an entire Go track, which I very much recommend that you,
and your listeners, uh, look up, uh, uh, learn about, uh, the website is goline.govlang.org.
So I don't know whether you remember when Java came out. Yeah. Java was supposed to be C++
done right. Yeah. Um, Go is C done right. Yeah. As a C programmer, I look at Go and it's like,
oh, yeah. This, this is how it should be. So give me an example. We had a horrible security
hold to do without DC, obviously handling. Um, it's been in the code for I think of a seven years.
Okay. Nobody noticed it. We found it with an internal audit. So as far as we know,
it's never been exploited, even though it was a root-level security hold. Yeah. But it's been in
the code for seven years. It happened because of a subtraction. Simply subtracting one number
from another. Yeah. And we got a range check wrong. Okay. That blew a hole in. Yeah.
Yeah. Completely, completely, completely come from my machine. One subtraction. The problem
with C and C is a wonderful language. I love it. I code it every day. You have to be perfect
and you have to be perfect all the time. Yeah. Nobody's perfect. Go doesn't have those problems.
I really, really encourage your listeners to check out Go. It's just a superb language. If I was
doing something again, I would probably do it in Go. I think. Okay. Um, it's, yeah. It doesn't have
those problems and yet it's still a systems programming language. So they're going to talk about
it tomorrow. I'd say there's a bunch of talks. Some of which talk about how Go is being used
at Google. Okay. In production. All for the listeners. All those, I've got links to those
recorded talks. Oh, cool. Very nice. Very nice. Yeah. I would, I would really recommend it.
Anyone who's, anyone who's used to doing C programming will look at it and go, yes,
this is what C++ should have been. Yeah. This is not my family. Sanity. I don't know. I'm not
that, I'm not a programmer. I just interviewed them for a little bit. It looked right. It looked
right. It looked beautiful. It looked clean. It looks elegant. It's, you know, it's the right thing.
I think it's the right thing to use. I'm what do you do for your spare time if you have a, um,
called Sanity? I have an eight-year-old. He takes up quite a bit of time. Um, you know, I do do
some recreational San by coding because it's still fun, you know. It's still, I just don't get as
much time as I used to. So why is it recreational to stuff that you're talking to do stuff and work?
Well, so I have stuff I have to do at work. And these days, Sanity is a very, very professional
product. So no piece of code, no code change gets into Sanity without two team engineer review,
without, you know, it's not as bad as the one. It's good. Well, you know, in terms of hostility,
but people hammer out the code. They make sure it's right. They make sure it's beautiful. You
know, it's done right. It's, it's professional programming work. So sometimes at the weekends,
it's just nice to sit down and write some, you know, write something new for fun. And then,
all right, it's still going to go through the review process when you've done it. But, you know,
pressure isn't on. Yeah. It's like some last one I did was some stupid bug where old
DOS clients, and I'm talking DOS-y, a pre windows, right? When we moved to 64-bit machines,
old DOS clients stopped working. And, you know, I've known about that bug and I knew what it was.
And I haven't, because DOS clients, who cares, right? And I, you know, one weekend,
then I saw, oh, screw it. I'm going to fix this. It'll be fun. Yeah. I know what to do. You know,
and it took me the weekend, but... And you know that somewhere, and some factory somewhere,
somebody is going, yes, that's right. Oh, oh, yeah. No, I knew the poor customer. Oh, my god,
thank you, thank you. You know, I know, you know, nobody could, nobody could make anybody do that.
I was done purely for fun, you know. If you're ever in the line for which it's called, you know.
Yeah. You mentioned earlier about the cloud. Nobody keeps, nobody keeps anything now anymore,
everything's in the cloud. So, obviously, we have no need for somebody. Probably true.
I'm not going to argue. I gave a talk at the SNES Storage Network Industry Association
about five years ago, saying, you know, the death of SMB. And things are moving to the cloud.
And the, essentially, application, nobody writes application protocols that depend on precise
local semantics running remotely anymore. If they do, they've written a bad program. So, my
argument is that if anyone, if a student, you know, Microsoft Access, you know, the way it has
multiple accesses to the database using file locking locally, if a student came up with the
design from Microsoft Access these days, you'd fail them because that's not how we do things now.
And, you know, in some ways, that's true. Sandry essentially is implementing a legacy protocol.
SMB3 has some very nice and interesting features that will extend its life. But eventually,
people, and you see this with Amazon, you see this with Google, you see this, you know, with many of
those apps, they're writing their own custom protocols to do exactly what they need and no more
against cloud, you know, against the storage that they're using. They're just not using perfect
file semantics on a LAN anymore. So, you know, in some ways, yes, Sandry's legacy, at least the
file sharing part of it, probably not the actual directory piece. I still find it fun. I still
think it's an interesting topic. But I'm not going to stand there and say, oh no, we're, you know,
some type of data. Yeah, I still have fun with it. You know, maybe eventually we'll move
under something else. But, I mean, even as a legacy technology, it's going to be around a long time.
You know, this thing's take a long time to move. So, you're a very unspoken
proponent of the GPL3. Yeah. I think it's a better license than V2. I actually have a talk
that you can link to and I can't remember is that URL as to why GPL3 is a better license.
And this is the funny thing that people don't understand. GPLV3 is actually a better license
for working with proprietary software than GPLV2. It actually is, if you want to link
GPL and proprietary software, GPLV3 is a license that you want your GPL code to be under.
And there's many reasons for that. You can read them in my talk. The only thing that
GPLV3 really stops you from doing is locking things down tight with DRM. So, if you want to use DRM,
GPLV3 is a nightmare. It's not a license for you. But then again, you know, screw them. I don't care
about DRM. I want DRM to go away. So, I don't really care that it's hard to use with DRM,
but for everything else, GPLV3 is a lot easier to use and to link with proprietary software.
We have many vendors who are selling legacy, you know, legacy fast sharing,
sambar to cloud gateway products, right? And they're much happier with GPLV3 because
it's actually easier to use with proprietary products. So, like I said, you need to read the,
I'll do a whole talk on it for exactly why I put a link into that. You have an age-year-old,
how do you or do you bother entertaining or trying to get your age-year-old into tech,
or do you push that, or how do you hack an age-year-old? I don't really. I show him interesting things
and let him play with what he wants. So, he loves LEGO. So, right now, we're doing a lot of LEGO.
So, I don't think you can make an age-year-old interested in anything.
You're certainly not computers, you know. The funny thing is, I think he will grow up in a world
where desktop PCs are really old people. One of the funny things I remember seeing,
and this was when he was much younger when he was about four or five, was, you know,
we sat in front of the desktop PC to play a game or whatever, and he started touching the screen,
because why wouldn't you? And then he was like, oh, oh, yeah, I'm like, no, you have to use it,
you know, oh man, yeah, it really was like, just got the scene with computer, yeah, oh,
that's quite how we used to do it. So, that's a change that's really coming.
So, yeah, no, I don't think you can. I think, you know, as long as they're having fun,
then they're learning. Right, I don't think we have anything else to cover in this interview.
All right. So, thank you very much for the talk, and tune in tomorrow for another exciting episode
of Hacker Public Radio.
You have been listening to Hacker Public Radio, where Hacker Public Radio does our.
We are a community podcast network that releases shows every weekday Monday through Friday.
Today's show, like all our shows, was contributed by a HBR listener like yourself.
If you ever consider recording a podcast, then visit our website to find out how easy it really is.
Hacker Public Radio was founded by the Digital Dark Pound and the International Computer Club.
HBR is funded by the Binary Revolution at binrev.com. All binrev projects are proudly sponsored
by Lina Pages. From shared hosting to custom private clouds, go to LinaPages.com for all your hosting
needs. Unless otherwise stasis, today's show is released under a creative comments,
attribution, share a like, and do those own license.