AV-Portal 3.23.2 (82e6d442014116effb30fa56eb6dcabdede8ee7f)


Video in TIB AV-Portal: FreeBSD

Formal Metadata

A brief history of the early days
Title of Series
CC Attribution 3.0 Unported:
You are free to use, adapt and copy, distribute and transmit the work or content in adapted or unchanged form for any legal purpose as long as the work is attributed to the author in the manner specified by the author or licensor.
Release Date

Content Metadata

Subject Area
Rod Grimes, involved with early FreeBSD, will give a brief history of the early days. Rod will talk about the time from the pre-patch-kit days up to the 2.x releases, with some mysterious history gleaned from the repo logs.
Group action Building Code Multiplication sign Source code 1 (number) Set (mathematics) Disk read-and-write head Mereology Traverse (surveying) Medical imaging Mathematics Semiconductor memory Different (Kate Ryan album) Cuboid Damping Office suite Physical system Area Scripting language Email Theory of relativity Binary code Electronic mailing list Thermal expansion Complete metric space Band matrix Type theory Tower Hard disk drive CD-ROM Spacetime Point (geometry) Aliasing Server (computing) Sine Computer file Patch (Unix) Virtual machine Device driver Portable communications device Number Product (business) Revision control Goodness of fit Internetworking Computer hardware Gastropod shell Computing platform Software development kit MIDI Patch (Unix) Projective plane Line (geometry) Frame problem File Transfer Protocol Software Personal digital assistant Network topology Formal grammar Greatest common divisor
Group action Code Multiplication sign Decision theory System administrator Source code 1 (number) Set (mathematics) Shape (magazine) Different (Kate Ryan album) Core dump Repository (publishing) File system Position operator Physical system Area Email Gamma function Electronic mailing list Bit Repository (publishing) Hard disk drive output MiniDisc Right angle Figurate number Point (geometry) Game controller Supremum Computer file Sweep line algorithm Patch (Unix) Virtual machine Device driver Online help Login Code Number Revision control Selectivity (electronic) Software testing NL-vollständiges Problem Task (computing) Modem Alpha (investment) Software development kit Dependent and independent variables Information Projective plane Expert system Core dump Directory service SI-Einheiten File Transfer Protocol Network topology Routing Limit of a function
Point (geometry) Web page Group action Link (knot theory) Multiplication sign Patch (Unix) Image resolution Set (mathematics) Client (computing) Revision control Wiki Mechanism design Core dump Position operator Software development kit Theory of relativity Information Software developer Weight Projective plane Division (mathematics) Bit Lattice (order) Limit (category theory) Statute Revision control
Metre Implementation Computer file Code Multiplication sign Decision theory Patch (Unix) Tape drive Source code 1 (number) Mereology Number Revision control Mathematics Goodness of fit Insertion loss Repository (publishing) Gastropod shell Authorization Physical system Computer architecture Scripting language Block (periodic table) Software developer Code Sound effect Bit Line (geometry) Cache (computing) Data management Kernel (computing) Process (computing) Repository (publishing) Network topology Buffer solution Universe (mathematics) Revision control Video game Figurate number Window
Implementation Group action Statistics Divisor Wage labour Computer file Code Patch (Unix) Multiplication sign Transport Layer Security Source code 1 (number) Design by contract Event horizon Information technology consulting Power (physics) Software bug Product (business) Data conversion Software development kit Email Theory of relativity Key (cryptography) Projective plane Electronic mailing list Data storage device Physicalism Bit Directory service Lattice (order) Human migration Type theory Data mining Process (computing) Network topology Password System programming Hard disk drive Right angle Freeware
good morning everyone can the people in the back girl hear me okay let's trying to started um you can hear me okay up in the back row good okay my name is rod Grimes probably everyone in here knows who I am by now the talk is a brief history of some of the early time of freebsd some of it before freebsd even started up to about baby release one point five or two would be the time frame i want to ask if there's anybody in the room that was in the project before the end of 94 please raise your hand I want recognize some people I believe Doug you were Justin Warner before the end of 94 yeah yeah not at the beginning on you for these are something and so we've got a group of some of the very early people here unfortunately many of the early people are no longer with the project they moved on to other things for one thing or another the 386bsd patch kit came about because there was a the 386 release was out there at 0.1 there were lots of changes floating all over the internet nobody was really collecting them and you kind of went around and you grabbed dispatch from selling us on this patch from so and so when you tried to integrate those into a source tree you were running into conflicts because the same people were changing the same sets of code especially in the networking area or the drivers that kind of stuff so a gentleman named Terry Lambert started this idea well let's make a patch kit and collect all these patches together resolved out the conflicts between them and give a way for a person to maintain a source tree that included a large number of patches and the ability to back those patches out and put new ones in as the patch kit evolved it was somewhere it was several thousand lines of shell script and that was done for portability because we needed to be able to patch this system basically on any UNIX platform because you weren't necessarily putting your sources together on a hosted system it was self hosting at that time but build world and those types of things just didn't exist there were a lot of Meg files all the make files were there to build the individual binaries but there was no way to do a complete tree traversal and build the whole pack the whole system after the patch kit was around I think it was maybe nine months Jordan Hubbard was another name that you should all recognize and hopefully recognize approach along a creek cd-rom about making a CD of this product project and and Walnut Creek was was very receptacle to that idea and we decided there was a mutual agreement between a couple of people that yeah let's kind of go do this and see where it goes there wasn't any we didn't have any project we didn't have any formal we did we did we didn't even have formal mailing lists at that time and Walnut Creek went out and bought some hardware and to for us to have a host I had a t1 at that time which was pretty good bandwidth back in 1993 and they were gonna let us use it which was an even better idea for us so they went and bought some hardware so we can put up a 386bsd box running at their side well in those days it was hard to just go into a computer store and buy hardware that would run bsd you had to be pretty careful about what you bought and Jack felt a was walked through a install of 386bsd on the hardware and it ran for about six hours and fell over and started panicking it wouldn't run for more than three or four hours so that hardware was boxed up and shipped to me in Portland to fix and it actually is what turned out if I remember right it had a cash problem the cash chips in him weren't capable of keeping up with how hard we were pounding on and it was a matter of replacing all the cash chips with a higher speed cast ship sending that attended him back down a godess bootstrapped I had the person it beneficial because the machine was in my hands and being reconstructed and we were going to reload it from scratch I named it I have a history of skydiving and that's where the name freefall came from that's a lot of people have asked about that over the years about why that name was chosen and stuff - freefall I had other machines name ground rush and sky rush that have are also skydiving related names ground rush and sky rush which are two aspects that occur during skydiving if you're if you're go below pull altitude the earth appears to explode below you as you get closer and closer to the ground hopefully you do something before you get there or are wearing something that does something for you before you get there yes and I'm not sure who named that now I don't know if I named it or if that was in Jordan going look at what happens if you don't and in fact there were some other machines there because I'm trying to there's a physical image or a mental image in my head of the physical room where these machines were located they moved at one time they were in one room freefall said on one machine after I had gone down to Walnut Creek the FTP cd-rom was built which was probably a hostname I should have mentioned I built the first FTP dot cd-rom at Jack Ville T's request and that they had been hosting some small FTP sites on a Sun for and they wanted to do something bigger and part of that was one of the large mirrors was shutting down I can't remember who it was they wanted to pull pull occurred what I shouldn't said mirrors one of the large FTP sites was shutting down and they wanted to bring a copy of that over before it went away no I don't think it was him tell but it may have been anyway so they had this idea to bring it in and host it themselves so I built another machine called FTP GCD rom-com the two machines set in a rather large room side by side freefall wasn't anything more than a mmm mid tower size case it has I think one four gigabytes cozy hard drive in it I believe the very first one we only had four Meg of memory it may have had 16 it quickly got 16 those two machines later moved into another room due to expansion of walnut Creek's they needed more office space so we moved the freebsd projects in with the walnut creek server into a back server room and i'm trying i seem to remember that there were five machines in there one of them would have been walnut creek servers so there's there's four other machines which is ftp free-fall and i don't know if it was thud at that time or not and there may have been a machine called time cd-rom comm but those machines were some of the very early ones later hostnames Jordan when Jordan came over much later the his whiskers machine I think went into that same same arena Justin would know we didn't even have a project name at that time so the host name actually came before the project how we ended up with free in both is a mystery other than it may have tickled David dream its name the I want to say there was about a dozen of us that were pretty actively involved in the patch kit and the starting of this project and basically the idea that that we could pull together and do this was managed by a mailing list that Nate Williams did who's the third member of the the founding 3 at University of Montana and we had a little MIT mailing us it wasn't even major dome eyes it was a sin mail alias was just a list of the people in it and if we needed to add somebody to it we just added them and somebody somebody's mail started getting bounced they just got deleted it wasn't mm kind of fortunately I haven't been deleted for some of my mail bounces but they rarely last for more than a day so anyway this group of people kicked around a whole bunch of different ideas and I'm not gonna go over any of the names cuz I can't even remember all of them the one that stuck we all know today's freebsd was picked by david dream men the group basically we voted on a list of names and that was the one that everybody likes the most and so you can blame our name on david griemann we
didn't have a repository at that point so we some tool selection I think had been done the whether the name was picked before the repository was created or not I'm not real clear on a lot was going on at the same time so anyway so creation the repository involved taking all of the 386bsd sources that were distributed by Bill Joel that's on a couple of different berkeley mirrors and somebody else had a copy of them anyway those were downloaded expanded out into the original ones were chunked up into floppy sized pieces this is the days of 14 4 and 28 8 modems so you didn't want to try and download 8 megabytes the source code all at once so anyway I had I had the wonderful task that somebody else picked the tools I don't think I had much input to it and was gone okay we're gonna put we're gonna do our source code control in CVS and so I went ok I got to go figure out what this thing that's called CVS is because I've never seen it before I was familiar with some of the predecessors SEC s and the RCS stuff that was used at the University so I had to go quickly grab the CVS back and read through all of that and figure out how to put a repository together and then one there were several people in the projects that had worked with CVS somebody pointed me to a set of tools that create something called the CBS route administrative directory that gives you things like okay we can decide who gets to do commits and we can produce these really nice mail logs majordomo got dropped in my lap and I'm not even sure how quickly that came in I think it was pretty early in the project probably before commits I'm not even sure if the very first commits went out in majordomo mailing list we had to have some way to manage our mailing list it had been up at Montana and moved those to Walnut Creek but the initial codes were imported and then I had I went through the patch kit which at that time I believe was 70 some patches and brought those patches in one at a time so they were all in isolated individual commits with all of the information that we had about who had developed them what they were changing why they were changing it that kind of stuff so we had some history to start with and once that was done we figured out well okay now we've got to be able to make it so that everybody can get to this stuff and so we started adding accounts to free fall so that literally you did all of your you're committing work on free fall many people that compiles and test builds over there and stuff you mean you don't want to you don't want to haul it off to your machine and then have the rest of the tree evolving outside of you and you're trying to get this just this little piece of code fixed we cook we quickly eight through four gigabyte hard drives it was it was pretty easy to do and that's kind of how the repository started going in and it was it was available I'm trying to think if I can't even remember how we distributed it first I don't think we had some CVS up wasn't available at that point in time I think a lot of guys just checked out first reasoned and tore them up and downloaded him CTM came CTM came in post one release i'm predominantly talking about this is before 1.0 and we don't have a release yet don't suffice yeah nan yes we their CVs up in there sup and we did use just sup in the very very early days but I don't know that we had at day one or I shouldn't say day one I should say in the first six months Turin FTP we didn't we I don't it was we never did it with our disk unless it was done individually something I mean there's no reason that somebody can't set up their own are distant and pointed at it and send it so it probably came in pretty early Julian else sure was a big help in that because he was already somewhere where they were using so I have the ominous task of being the person responsible for the core team there's not very much documentation on this piece of history I was in a physician at Walnut Creek Jordan was in Ireland working for Lotus yes and I'm not at in 1992 I was not a technical expert on all of BSD I had worked in many operating systems internals I had a pretty good footing in operating systems but I didn't know the vm system like the back of my hand so I needed people that did as technical advisors to what we were trying to do and to make sure that we didn't do stupid things Kirk was at the University of California and wasn't really available to us he was still doing CSR G work we needed file system experts we needed a device driver experts we needed user Lance I'm not a user land person as much as I've done in user land we needed people that were experts in many many different areas so I just started collecting people some of them came very willing to the idea of the core team some of them didn't took a lot of persuasion to go look you know we really really need you some of them had con with other things going on precluded them from becoming thinking they say yeah I'll work with the project but no I don't think it would be a good idea if I sat on the core team of it there may be conflicts with my employer there may be conflicts with other things they were doing I can't even remember what the exact size of the initial core team was I think it was nine and it's still nine I don't know has that number changed over time Justin okay it was I'm remembering about eight or nine people that were were there so we went on to spend I'm not even sure what the exact timeframe is I believe it was April or May of 93 when I went down to Walnut Creek Sitaram and I don't think it was till December that we spun the 1.0 alpha very very first copy of FreeBSD so I have the wonderful chore of creating the release engineering position because that's it wasn't there wasn't even a position it was just work that I had to do I was a janitor you know I had to kind of sweep all the bits out the door we went through a rather long series of alpha-beta I believe there was a delta and a gamma of the 1.0 I think huh is that the four huh yeah and I believe there was a Delta that didn't it didn't live very long there was an epsilon - but I don't huh I don't you know the and these weren't these were test releases immediately just making sure that we had bits that everybody could run and and had a way to get them to people somewhere and I recently found it I have the what's called the golden copy of the final one dot oh release it was actually sent out from Walnut Creek for pressing it's that time we couldn't even we couldn't master a CD I had to haul all I had to pack all the bits up once we had him in a shape that we liked and take them over to his son for and Bob Bruce actually did the burning of the very first what they call a one-off that lead I don't know whether it's it was unfortunate for Bill Jellison that he couldn't attend this conference called nlog I'm gonna call this the zeroth bsd conference or at least the zeroth FreeBSD conference and that because of his cancellation that n log someone in the Netherlands this is another one who's Linux users group I think I added one too many use in there they they huh it's right is it right okay thank you I was a UNIX users group okay the the they contacted Jordan about well we've got this slot open we need somebody to come present so Jordan contacted me and goes look we're going we're going to end log and we're going to present about FreeBSD and it's going to be me and you and those were decisions Jordan just we're gonna go do this okay I fell okay that sounds good I didn't even have a passport at that time I think the conference was thirty or forty five days away we were getting pretty close I had to run over in San Francisco to the consulate and file to get a passport in those days it was much easier and they were able to give me a passport pretty quickly
well that was anyway we went to n log some of the very early core team members and developers were at n log and I'm going to try and hit these names I may not remember all of the people that were there that were heavily involved in the project I was there Jordan Hubbard Paul Helen Camp Widow ban Rho G and Paul Richards I know we're all of their Kirk McKusick was there and Linus Starr basalt was there and we actually mean Jordan and Kirk went out to a dinner that night with Linus and some of the N log members and that was the first time I ever had met Kirk having been following his work for 10 or 15 years 10 years at least and it was the first time I met Jordan I flew into the Netherlands and and trained over to Utrecht and met Jordan an hour before we were going to give a joint talk that's also the last time I spoke in front of a crowd to the users group of this size that's literally the last VSD related conference that I've attended so I've been missing for 24 years from the conference arena at the same time we started the FreeBSD project Christian meet rescue and a few others at the University had started up the net bsd project i think they actually they got named and started a month or two a month before us they were kind of they started as as soon as 386bsd was out I'm pretty sure that they they started that project Kirk may know I can't remember didn't that didn't did net BSD start the project after Bill hadn't released 386bsd or were they already started it was right after yeah I think I think they actually started while we were high well that was the same reason that we started I mean there was the patch kit had been created as an external mechanism to try and deal with some of it but yeah okay anyway the there was a large feeling in the community that this may not be the best thing for the futures of the bsts to have two different Forks that were basically doing the same work I was actually a very early contributor to net BSD because of the patch kit stuff because they picked that stuff up and used it too and so there were several meetings between me and Kristin meets rescue and I don't know if I'm saying that name even close Kirk you would know to meet you you meet you okay to try and merge the two teams and not have to separate projects and unfortunately that has met with failure on at least the two attempts I made at it and I believe there was a later a much much later attempt to merge the teams that has not done I think in the long run this has been healthy for the bsd community to have multiple groups because it allows it allows diversity in what's going on it allows some it allows division of conflict in personalities which was some of the problems we had in the early days I am NOT an easy person to get along with you never ever want to work for me you just don't so we those fails then we also had a
little problem this is a fairly famous problem that USL novella started to BSE and we'll just call it the BST lawsuit because they sued CS or GE they sued bsdi they sent letters to Walnut Creek senior um for some reason they picked my name and they sent me a letter and and so we had this this little problem and there were some tainted bits that they didn't want out their client their initial claims were much larger they final resolution on it actually ended up being fortunately a very small set of bits and so an agreement was reached and for ten years that agreement was pretty secret I've just recently learned that agreement is now public record and that due to the statute of limitations and Public Information Act in fact it's yes that the University of California is a public entity somebody managed to get it out there so we can all read that agreement I would encourage you to read it because there's there's some interesting claims in it it's it is available on the if you go to the weekend and look for the BST lawsuit the wiki will point you off to yes yes Wikipedia this this is this is not yes not here well no I think is it okay song rock club but there's a link through their Wikipedia page to that points to the agreement yeah I don't anyway the the agreement was made and that put us in the position of what do
we do now we have we have a situation where we've done all this work in a repository but that repository is technically contaminated and there's different approaches to trying to fix that problem and that BSC's approach was to I believe addicts the files that were in in conflict and replaced them but that meant two files were now in the attic of their repository our decision was that okay we don't it's not easy for us just to remove the history on these files and we didn't want to do that because that creates a hole in the history we also knew that well okay for for light is going to be the replacement for the solution of this and I can't huh it's just more for like it's it's this it's the second release of for for without the encumbered bits it wasn't called light - was it Kurt yeah yeah that was later yeah I made the contacts and I think it's the very first time I ever spoke with Kirk - for Walnut Creek to get a eight millimeter tape of the four four light sources and that was simply done so that we would have one in a filing cabinet and go yes we've got that we've done we've done the proper paperwork with the university to have a valid copy of for for a life the tape was read once and the bitch pulled out and I'm not even sure that I I did that to verify that the tape was good that's not where I got the bits to import the depositor I just downloaded those it was to make sure we had the tape that got stuck in a cabinet that we could actually be used in the future if we had to so we had to read which is what we decided to do was reinforce the sources basically we picked four or four like up and brought those into the repository and there was the missing bits that had to be reimplemented so this gave me as the repository manager a time window that I could work on how to import these bits because the first time it was done fairly hastily and I hadn't had a lot of experience with CVS at that time so I spend a little more time somewhere there is a couple of hundred lines of shell script that actually did the import it was all just automated and that's because I could sit there and go through it and then look at how the repository come out now that's got problems here and go back and resend it while the development work was going on to replace the missing bits this creates some it's non accessible history but there are actually Oh most of the repository has two versions of 1.1.1 or I should say there are two versions of the 1.1.1 files which is what you get when you do an initial import they're in two different places there's two repositories there's the old CVS repository and there's the new and CVS repository so if you try and look at the sources from a 1.0 FreeBSD release you'll will find a whole bunch of 1.1.1 files and those may not match the sources that are on a three dot and was to two something I think is where we re implemented but anyway I know they won't match the ones on a3 where there were changes between four for BSD and four for BSD light so there'll be a small number of them that won't match but they're those one-one-one files are from two different places we lost some time reimplemented the bits of course the the I was talking with Kirk the other night trying to figure out exactly what it was that had to be reeling to remember part of the vm system and part of the block IO system that got X cured because of the way things had to be removed and most of that work was done by David dream and a gentleman named John Dyson and it had to be white roomed it was our first experience with having to do white room development David had been exposed to the source code so he had a pretty good idea of what was there and basically fed John with with api's and here's what we have externally so you go figure out how to implement something inside of here that works David had become pretty efficient proficient with how the vm system worked internally and i think they actually made some significant changes into from what we had before to what went back into the tree i'm not sure if that was the first unified buffer cache implementation that they did because I think they spun it at a later time in
some of that processing we did some things to the repository that are not so nice I went through there there was when we did NCBS there were some bits that we knew were clean that were in our prior work anything that had been written after the 386bsd 0.1 release by an outside party was clean because that wasn't it didn't it didn't have a CS RG lining so we brought all those bits over there were there was there were some code from external sources like did for for ship with patch we ran into this one we're looking to replace for that for some reason there's an import of patch onto a vendor brand externally in the very first days I'm still trying to figure out why maybe was there a Berkeley patch and a good new patch yeah I think that's what I think we act we am we imported the new patch because there was some features in it that we needed okay so it was an external it was available externally from so we could actually we could actually go back and grab all of that work and and bring it in and not have to worry about being encumbered other things that happened later because we had developed these habits though it's okay to go over here and grab this part of the tree and repository copy it over here and leave all the tags there I now should go to AMD and collect a very very large royalty check because evidently in 1993 I created the kernel sources for an AMD 64 architecture that we've we just found this the other night when we were looking at the repository and there's some I think it was John Baldwin it keeps going keeps going back in the tree and ends up at at 1540 as a version number when you're back walking almost anything in the kernel and you end up at this 1540 version and that's where the repository copy was done of the i386 bits to the AMD 64 bits and when you do a repository copy none of the dates get changed none of the tags you changed none of the authors get changed so I thought the tags were showing there though oh you rename the tags that's okay they're still there they're just renamed okay see how see how easy it is to mangle a repository the some of it some of the side if so some of the side effects of this is you can't check out things by date and getting me a tree that matches that date anymore because there's yeah yeah that's correct you do you do get all the files that were there and you do get some versions of the files that were there you just get a whole bunch of other stuff that wasn't the the may have not been written 20 years before half I think well I want to
thank all of you because you're the ones that have made this project last 25 years some special thanks to Michael Dexter who couldn't attend this year and that's why I'm here dan for keeping this conference put together and for Shanghai me into doing this I had three days to prepare and I wanted to socialize so I didn't do a very good job of preparing and and and and John Baldwin for the the dev summit work that made it possible for me to just kind of socialize with all of you
if there's any questions we've got we're out of I think we're right on time ending but Dan says I've got time if you guys have some questions Glen I noticed I did that reluctantly when you drag me back I think the challenge was already there Peter he's already been working on it Doug that's that's up for the powers with me huh why no it's easy to solve but it's up to some other powers it's not at my request I don't think it's fair that I short-circuit the process I did submit a PR two years ago that's been committed okay we're actually I'm supposed to work with Glenn later today to get my PGP keys back signed up so I can I haven't been able to log into freefall since sometime in 1998 I lost a hard drive that had my copies of the SSH keys so no it was later than that was 2001 because we went huh it may be gone but my home directory with my last source tree is still sitting there there's we looked at some password files I was trying to glean some history so that I wouldn't leave names out and what I decided to do is basically leave almost all of the names out it was actually easier the password file was done sequentially initially so that there was there was a way for me to know very well when people joined the project Jordan or Justin Gibbs joined the project the day after me according to the password file no way that probably didn't get picked up did it I believe that is Jordan's ID that he changed later it only makes sense any other questions no no this is it some some strange events that have happened that led up to this the there has been a UNIX related users group meeting in Portland Reagan which is where I'm from for I'm gonna say 27 or 28 years and I was involved with that in the early days it has recently been restarted by Andrew fresh he's sitting down here in the last year and a half and some how I had still been on the mailing list that had been there for the last five or six years Ted mental stat it was a committer at 1 a committer he has a free fall account so he must have been a commander at one time hosts the mailing list that was used for that and so I still got the mails because here's some interesting had Peter knowing that for the last 20 years he could have sent an email to our Grimes at FreeBSD org he would have but he didn't because he never thought to try it neither did Julian Elser who I Michael Dexter was at a conference that had to do with storage products in the Bay Area about two months ago maybe four months ago anyway Julian Elster was there who's another very very early project member he's a patch kit contributor I'm pretty sure he did our first Guzzi code implementation that anyway Michael called me and then put Julian on the phone and Julie's like well alright how do i email you I said what's our crime free beast you've got to be kidding me so if you know of people that you haven't seen around the project and you're trying to contact them you might just try their freebsd org mailing list most of us left longstanding forwards the unfortunate thing of mine is it gets about a thousand spams a day Doug oh we don't we don't have enough conference time to talk about that um I would have spent a lot more time with Justin having having known that that he was going to lead the foundation and be a figurehead would definitely be one thing anybody else you're gonna let me out of here this easy a little of this a little of that but as little as I possibly can in the technical arenas I make gainful income operating an executor bader for a gentleman I do some other physical type out work I do i fault I used to treat I treat opted for three years for a guy I do tree falling still on contract laborers I do what's called precision tree falling where you're you know two feet down the side of the house don't miss so far I've been pretty lucky mistakes are a lot more expensive because I can't I can't just redo it okay there we go new host any others I do do some contracting and consulting work some of it has been freebsd oriented I'm still believe it or not I've actually been there in the background the whole time I'm still running it I have my own private lab and cluster and stuff and I have followed it from the background I look at each release I probably downloaded every release or at least every other release since 1.0 the glands looking at me with a growl on his face because I didn't I didn't bother to point out some small anomalies and stuff that actually turned out to be really big headaches yes it's that's kind of what I want to says it's not really that big of a deal it causes me a little grief and just dealt with it and went on Kirk I don't know if it's what I like the least I was not happy with the conversion to SVM that I'm sorry guys but that I'm sorry I didn't the the the gain versus pain factor for me was not very good when that conversion happened and I was okay as long as CVS was supported us as a it was the bits were still being shoved back into there but when that got cut off I had a huge pain factor but that I can understand oh I can understand why it had to happen it's reasonable so that would be the the thing I disliked the most I don't know if that's really added to the code base itself or or not and probably the thing that I it's unfortunate mmm you're asking me to dig through a lot of time 20 years I can tell you what I'm excited about most recently and that's the implementation above a hypervisor beehive and seeing that code come in for me personally it gives me some roads to go down though I haven't been able to I run my cluster runs on ESXi so I would just love to go and swap swap in be - I can't right now it's getting close if we could get nested hosting working it would be a big boon because then I could just look yes ex I under be - and my cluster would be on beehive and then the migration becomes easier ZFS is a big win I don't use it a lot but from a technical aspect I'm really happy to see that code I guess that's last chance up in the corner Michael Lucas which one and how many times Glenn again you're welcome it's it's really you and people like Dan it's our conferences and John that have kept it going all's I did was start the wheels rolling sit we're done