DragonflyBSD专访: Jeffrey Hus (上)

DragonFly BSD论坛
dragonflybsd
帖子: 20
注册时间: 10 5月 2016, 20:20

DragonflyBSD专访: Jeffrey Hus (上)

帖子dragonflybsd » 29 7月 2016, 22:18

本期访谈,DragonflyBSD微信公众号邀请到FreeBSD最早的10个committer之一Jeffrey Hsu,在此次访谈中,Jeffrey从BSD的发展历史、个人经历、学习经验等方面进行了分享,其中不乏很多BSD发展中的有趣小故事。下面就随着小编一起进入此次的访谈吧。

此次访谈将以连载的方式在9月份的推送中陆续放出。

We are the product of our nature and ourenvironment. If I had gone to MIT, I
have no doubt I would have become a LISP programmer specializing in AI.
Instead, I went to UC Berkeley in the late 80's at a time when 4.2 and 4.3 BSDUNIX was being developed there, so I became a C and Unix programmer.

我们是天性和环境的产物,当年如果我去了MIT,我可能会成为一名人工智能的LISP程序员,但是我在80年代末期去了UC Berkeley(伯克利大学),当时那里正在进行4.2 BSD和4.3 BSD的开发项目,我参与到这些项目的开发中,并成长为了一名UNIX的C程序员。

The group that was developing BSD UNIX,CSRG, was located on the 4th floor of Evans Hall. It was a small group consisting of MikeKarels,Kirk McKusick,
and Keith Bostic working on 4.3BSD. Bill Joy had already left for Sun
Microsystems, afteradding a Virtual Memory facility to SysV UNIX for the
VAX in 4.2BSD. Even though the group was small, because it was a
source code and not binary-only distribution, it has hundreds ofcontributors
from its user base.Unbekownst to me, Jordan Hubbard wasworking as a Sys
Admin on the 2nd floor of Evans Hall. Jordan and I were at Berkeley at the same time --- the same building even --- but we wouldn't meet in personuntil
many years after that.

当时开发BSD的CSRG在Evans Halls的4楼。当时CSRG的人不多,主要是MikeKarels,Kirk McKusick和Keith Bostic三位研究员在进行开发(从事BSD系统最早的几个开发者)。Bill Joy在VAX上使用4.2 BSD给SysV UNIX实现完虚存系统后就去了Sun Microsystems。虽然CSRG规模不大,但是因为BSD的代码是开源的,在BSD的用户中有上百个代码的贡献者。当时Jordan Hubbard(苹果 OS X 和 IOS 操作系统项目的领导者,FreeBSD的创始人之一)在Evans Halls的2楼当系统管理员。虽然Jordan和我在那段时间里一直都在UC Berkeley的同一栋楼里工作,但是我们在多年之后才真正的见面。

UNIX was the intellectual property ofAT&T, which gave free licenses to
universities. But everyone prefered BSD to SysV, so the first thing a university would do on getting itsAT&T license was send in $100 to get a BSD Unix tape from CSRG. By the time 4.4BSD came out, most of SysV UNIX had been
rewritten except for a few files,which was to be the subject of the infamous
lawsuit between AT&T and UC Berkeley, the delay from which leaddirectly to
the creation and rise of Linux.

当时AT&T拥有UNIX的知识产权,并且免费提供license给大学使用。但是当时大家喜欢BSD胜过SysV UNIX,因此大学一般会在得到AT&T的UNIX免费许可后,寄100美元给CSRG来换取BSD的版本发行磁带。在4.4 BSD发布时,除个别文件外其他的SysV UNIX代码都已经被重新写过了。但正是因为那几个没被重写的文件就成了AT&T和UC Berkeley官司的导火索,也正是这场官司促进了Linux的发展(因为此官司的影响BSD无法继开发,很多BSD的开发者转而从事Linux的开发)。

There was a famous terminal room in thebasement of Evans Halls where the CS majors hung out calledB50. At that time, PDP-11 and VAX microcomputers cost hundreds ofthousands of dollars and were huge time-shared departmental machines. The CS department had donated a VAX-11/730, named Miro,
for theundergrads to play with.We were on the ArpaNet and had email, but no
Internet or WWW. The CS machines at Berkeley were named after famous
Renaissance artists.The source code for BSD Unix was on amachine called
Dali. So the first thing a budding kernel hacker would do is sign up for a
source account on Dali and read the BSD code. Compiles took hours and we could only reboot the machine lateat night, as it wastime-shared during the
day.

当时在Evans Halls的地下室有个有名的实验室叫B50,大多数读计算机的学生都在那里出没。但是那时PDP-11和VAX的计算机都非常的昂贵,不同的系的只能在固定的时段使用这些计算机。计算机系捐赠了一台VAX-11/730(命名为Miro)给计算机的本科生使用。我们当时使用ArpaNet并且拥有电子邮件,但是没有Internet访问和WWW。当时在UC Berkeley,计算机系的主机都以文艺复兴时期的艺术家命名,BSD的源代码在一台叫做Dali的计算机上!所以新来的内核程序员要做的第一件事情就是在Dali上申请一个获取代码的帐号并阅读BSD代码。因为在白天我们和其他系共用这些计算机。而编译BSD要花好几个小时,我们只能在晚上才能重启计算机,
It is often said that naming is one of thecentral problems in Computer
Science. My version of that is slightlydifferent. To this day, much of what I
know about art canbe attributed to the CS machine names. I would go to
museums like the Chicago Art Institute, see painting by famous artistslike
Degas or Monet or Matisse, but to me, they were machines Ihad played with in my undergraduate days!

大家通常认为命名是计算机科学中的一个关键问题,但我的看法略有不同:我对艺术的了解完全源自于那些UC Berkeley计算机系的主机名!要不是因为这些主机名,我可能会到芝加哥艺术中心去看看Degas,Monet和Matisse的画。但在当时,这些“艺术家”不过是我在本科时使用的主机罢了。

As an interesting aside, one night we saw anew guy with beard sitting in front of one of the terminals inB50. None ofus recognized him. He smiled, but
didn't really talk to anyonein the room. We never saw him again, until
graduation, when it was revealed that Steve Wozniak, years after Apple
hadgone public, had come back incognito under an assumed alias to finish his CS degree at UC Berkeleyand only a few professors knew about hisreal
identity. In hindsight,that was the Woz we saw that night in B50!

顺便说一个在B50的小故事。有一天晚上,我们在B50看到一个大胡子新人,没人认识他,他很友好的对大家笑了笑,但是从来不和任何人说话。那次之后我们就再也没有见过这个大胡子新人。直到毕业时我们才了解到,那个人是Steve
Wozniak(苹果的创始人之一)。他在Apple上市后匿名回到UC Berkeley来完成他的计算机科学学位,只有为数不多的教授知道他的真实身份。

With the advent of the Personal Computer inthe late 80's, we could suddenly
buy our own dedicated computer for$3000 --- no small sum of money, but
affordable for a dedicated enthusiast --- and experiment with 386BSD on it. Around the same time, two different groups,FreeBSD and NetBSD, forked off
386BD and were providing open-source versions of BSD UNIX that ran on the PC. Since not much was known about either distribution at that early stage, I somewhat arbitrarily picked one, FreeBSD, at the time because it had just
released a versionat the time I bought my $3000 PC and I wanted the latest
version. The lesson here for startups andopen source distributions competing for early market share is to release early and often. Amazingly, with this combination of a personal computer and open source BSD,it was now possible for anyone to do kernel work with his own resources, without needing
departmental backing.

在80年代末期,个人电脑开始出现。一台个人电脑需要花费3000美金,这在当时是笔不小的开销。但是对我们这些计算机系统开发的狂热分子来说,3000美金买一台专属计算机完全值得!搞一台装上386BSD!与此同时FreeBSD和NetBSD开始出现。在刚开始大家对FreeBSD和NetBSD都不太了解,当时我选择FreeBSD是因为在我买计算机时FreeBSD正好有一个新版本,我要在我的计算机上安装最新版的BSD!在这个事情上我们可以获得一个不错的经验:对于初创公司和开源软件的发行版本来说,尽可能早的和尽可能多的发行新版本对抢占先期市场会非常有利。随着个人电脑的出现和BSD开放源代码,在那个年代任何人都可以从事内核开发而不再需要庞大的资金支持(比如大学的计算机系)。

Nowadays, there is a lot of open-source code out on the Internet. But at that
time, most software was commercialand only binary executable or libraries were provided. As a programmer, I've always said, just give me the soource
code and I can doanything. So that early trend popularized by BSD Unix
providing source code lead the way to much of the richness we have now in
the open source community today.

现在的Internet上到处都是开源软件代码。但是在FreeBSD刚出现的年代,大多数软件是商业软件,只提供可执行程序或者函数库。作为一个程序员,我总是告诉别人:“只要给我源代码,我可以改出你想要的任何东西”。现在开源社区有如此种类丰富的软件其实可以追溯回当时BSD开放源代码所引领的开源风潮。

回到 “DragonFly BSD论坛”

在线用户

用户浏览此论坛: 没有注册用户 和 1 访客