前微软老员工讲为什么 bing 不如 google# Programming - 葵花宝典
z*r
1 楼
用http://www.buckeyeinstitute.org/higher-ed查Ohio州的公立大学教授的工资，不知道列出来的数据是9个月的还是12个月的？感觉9个月的话有点高了，12个月的话有的教授工资又有点偏低了。非常感谢！
A*i
2 楼
直接说是因为politics, 写得可以看出来带点情绪， 可能当时确实被politics阴得不
轻。
链接：
https://www.quora.com/Why-arent-the-engineers-at-Microsoft-able-to-make-Bing
-better-than-Google-even-though-the-employees-at-both-companies-are-equally-
smart/answer/Bartosz-Milewski?share=1
正文：
I worked at Microsoft in the early 90s as the project lead for the content
index, so I know exactly why Microsoft failed in this area. The reason was
company politics!
Content index was the first step towards the internet search engine. Since
the internet was in its infancy, our goal was limited to creating a search
engine for the local disk on a PC. In 1995, when I left the company, we had
a working Windows Search engine, which became part of Windows XP and which
is part of Windows to this day. In fact, we implemented a distributed engine
that could index several machines. Of course, scaling this thing to the
whole internet would be a non-trivial task; but remember, these were still
the early nineties, and we had a head start. If the management weren't so
short-sighted, we would have gotten there before Google was even established
in the late 90s.
So what happened? The content index was part of project Cairo, led by Jim
Allchin. It was a grandiose project that would provide a new windows shell,
a new file system (OFS), and a content index. The first gross mistake was
Allchin's idea of putting the content index in the kernel of the operating
system. As a project lead I decided to hedge our bets and create an
alternative implementation in user space -- ostensibly for testing, since
the OFS was not ready yet. I had to defend this decision at every management
meeting, but it worked. The index had quickly became indispensable for the
whole XP team, successfully indexing all their sources. (We actually had a
feature that would let you find the definition of a function, or a data
structure; or all its uses.)
The Windows XP team, led by Dave Cutler, was very much against Cairo,
because they felt is was encroaching on their territory. There was some kind
of competition or war between Allchin and Cutler, which culminated in a
deal: Allchin agreed to cancel project Cairo and was rewarded with a
promotion to Senior Vice President. Don't ask me how it was possible. In a
healthy company Allchin would have been fired.
So one day, we were told to abandon the content index and find a new project
. That really pissed me off, so I asked for a meeting with Cutler, Allchin,
and Mark Zbikowski, who was the mangager of OFS and my immediate superior. I
told them that we had a working version of the content index, in user space
, that had no dependency on OFS or anything else in Cairo. They let the
content index stay, but soon after I was let go. The index is still there,
but it's been crippled, its development stunted, and it never had a chance
to develop into an internet search engine.
轻。
链接：
https://www.quora.com/Why-arent-the-engineers-at-Microsoft-able-to-make-Bing
-better-than-Google-even-though-the-employees-at-both-companies-are-equally-
smart/answer/Bartosz-Milewski?share=1
正文：
I worked at Microsoft in the early 90s as the project lead for the content
index, so I know exactly why Microsoft failed in this area. The reason was
company politics!
Content index was the first step towards the internet search engine. Since
the internet was in its infancy, our goal was limited to creating a search
engine for the local disk on a PC. In 1995, when I left the company, we had
a working Windows Search engine, which became part of Windows XP and which
is part of Windows to this day. In fact, we implemented a distributed engine
that could index several machines. Of course, scaling this thing to the
whole internet would be a non-trivial task; but remember, these were still
the early nineties, and we had a head start. If the management weren't so
short-sighted, we would have gotten there before Google was even established
in the late 90s.
So what happened? The content index was part of project Cairo, led by Jim
Allchin. It was a grandiose project that would provide a new windows shell,
a new file system (OFS), and a content index. The first gross mistake was
Allchin's idea of putting the content index in the kernel of the operating
system. As a project lead I decided to hedge our bets and create an
alternative implementation in user space -- ostensibly for testing, since
the OFS was not ready yet. I had to defend this decision at every management
meeting, but it worked. The index had quickly became indispensable for the
whole XP team, successfully indexing all their sources. (We actually had a
feature that would let you find the definition of a function, or a data
structure; or all its uses.)
The Windows XP team, led by Dave Cutler, was very much against Cairo,
because they felt is was encroaching on their territory. There was some kind
of competition or war between Allchin and Cutler, which culminated in a
deal: Allchin agreed to cancel project Cairo and was rewarded with a
promotion to Senior Vice President. Don't ask me how it was possible. In a
healthy company Allchin would have been fired.
So one day, we were told to abandon the content index and find a new project
. That really pissed me off, so I asked for a meeting with Cutler, Allchin,
and Mark Zbikowski, who was the mangager of OFS and my immediate superior. I
told them that we had a working version of the content index, in user space
, that had no dependency on OFS or anything else in Cairo. They let the
content index stay, but soon after I was let go. The index is still there,
but it's been crippled, its development stunted, and it never had a chance
to develop into an internet search engine.
l*t
3 楼
12个月
【在 z****r 的大作中提到】
: 用http://www.buckeyeinstitute.org/higher-ed查Ohio州的公立大学教授的工资，不知道列出来的数据是9个月的还是12个月的？感觉9个月的话有点高了，12个月的话有的教授工资又有点偏低了。非常感谢！
【在 z****r 的大作中提到】
: 用http://www.buckeyeinstitute.org/higher-ed查Ohio州的公立大学教授的工资，不知道列出来的数据是9个月的还是12个月的？感觉9个月的话有点高了，12个月的话有的教授工资又有点偏低了。非常感谢！
ET
4 楼
you need to read one of immediate comment of this answer. Based on that
answer, what he did is not relevant to why Google is better.
Bing
equally-
【在 A********i 的大作中提到】
: 直接说是因为politics, 写得可以看出来带点情绪， 可能当时确实被politics阴得不
: 轻。
: 链接：
: https://www.quora.com/Why-arent-the-engineers-at-Microsoft-able-to-make-Bing
: -better-than-Google-even-though-the-employees-at-both-companies-are-equally-
: smart/answer/Bartosz-Milewski?share=1
: 正文：
: I worked at Microsoft in the early 90s as the project lead for the content
: index, so I know exactly why Microsoft failed in this area. The reason was
: company politics!
answer, what he did is not relevant to why Google is better.
Bing
equally-
【在 A********i 的大作中提到】
: 直接说是因为politics, 写得可以看出来带点情绪， 可能当时确实被politics阴得不
: 轻。
: 链接：
: https://www.quora.com/Why-arent-the-engineers-at-Microsoft-able-to-make-Bing
: -better-than-Google-even-though-the-employees-at-both-companies-are-equally-
: smart/answer/Bartosz-Milewski?share=1
: 正文：
: I worked at Microsoft in the early 90s as the project lead for the content
: index, so I know exactly why Microsoft failed in this area. The reason was
: company politics!
a*o
5 楼
不会吧，9个月的吧，12个月才这么点。。。
【在 l*******t 的大作中提到】
: 12个月
【在 l*******t 的大作中提到】
: 12个月
A*i
6 楼
well, I read those comments before I started this post, you may notice that
there are also several other answers under the same question
"Why aren't the engineers at Microsoft able to make Bing better than Google
even though the employees at both companies are equally smart?"
, those questions also provide some technical reasons like the comments you
mentioned. The point in this guy's answer is that Microsoft had a big
advantage of developing a search engine in the early stages of the internet,
and as mentioned in this guy's answer, Microsoft had already have a working
search engine on windows systems, but because of politics, this search
engine never got a chance to go further.
【在 ET 的大作中提到】
: you need to read one of immediate comment of this answer. Based on that
: answer, what he did is not relevant to why Google is better.
:
: Bing
: equally-
there are also several other answers under the same question
"Why aren't the engineers at Microsoft able to make Bing better than Google
even though the employees at both companies are equally smart?"
, those questions also provide some technical reasons like the comments you
mentioned. The point in this guy's answer is that Microsoft had a big
advantage of developing a search engine in the early stages of the internet,
and as mentioned in this guy's answer, Microsoft had already have a working
search engine on windows systems, but because of politics, this search
engine never got a chance to go further.
【在 ET 的大作中提到】
: you need to read one of immediate comment of this answer. Based on that
: answer, what he did is not relevant to why Google is better.
:
: Bing
: equally-
l*s
7 楼
我帮你查了一下
是实际收入
有的是9个月的，有的是10个月的，有的是11个月的，有的是12个月的
我的是11个月的
【在 z****r 的大作中提到】
: 用http://www.buckeyeinstitute.org/higher-ed查Ohio州的公立大学教授的工资，不知道列出来的数据是9个月的还是12个月的？感觉9个月的话有点高了，12个月的话有的教授工资又有点偏低了。非常感谢！
是实际收入
有的是9个月的，有的是10个月的，有的是11个月的，有的是12个月的
我的是11个月的
【在 z****r 的大作中提到】
: 用http://www.buckeyeinstitute.org/higher-ed查Ohio州的公立大学教授的工资，不知道列出来的数据是9个月的还是12个月的？感觉9个月的话有点高了，12个月的话有的教授工资又有点偏低了。非常感谢！
i*h
8 楼
一厢情愿的说法
他这个目录和狗搜索还差好多啊
that
Google
you
internet,
working
【在 A********i 的大作中提到】
: well, I read those comments before I started this post, you may notice that
: there are also several other answers under the same question
: "Why aren't the engineers at Microsoft able to make Bing better than Google
: even though the employees at both companies are equally smart?"
: , those questions also provide some technical reasons like the comments you
: mentioned. The point in this guy's answer is that Microsoft had a big
: advantage of developing a search engine in the early stages of the internet,
: and as mentioned in this guy's answer, Microsoft had already have a working
: search engine on windows systems, but because of politics, this search
: engine never got a chance to go further.
他这个目录和狗搜索还差好多啊
that
you
internet,
working
【在 A********i 的大作中提到】
: well, I read those comments before I started this post, you may notice that
: there are also several other answers under the same question
: "Why aren't the engineers at Microsoft able to make Bing better than Google
: even though the employees at both companies are equally smart?"
: , those questions also provide some technical reasons like the comments you
: mentioned. The point in this guy's answer is that Microsoft had a big
: advantage of developing a search engine in the early stages of the internet,
: and as mentioned in this guy's answer, Microsoft had already have a working
: search engine on windows systems, but because of politics, this search
: engine never got a chance to go further.
z*r
9 楼
非常感谢！faculty的工资看来确实比工业界的工资要低阿。
【在 l*******s 的大作中提到】
: 我帮你查了一下
: 是实际收入
: 有的是9个月的，有的是10个月的，有的是11个月的，有的是12个月的
: 我的是11个月的
【在 l*******s 的大作中提到】
: 我帮你查了一下
: 是实际收入
: 有的是9个月的，有的是10个月的，有的是11个月的，有的是12个月的
: 我的是11个月的
l*s
10 楼
He thought "Of course, scaling this thing to the whole internet would be a
non-trivial task;"
但是在那个时代，微软整体上是在Internet的步伐上落后，甚至有意打压拖延的，即使
简单的Scale的想法也没人愿意去予以实现。这不仅是一个小的政治性的斗争的问题，
最多算是当时微软对于互联网开发缺乏预见性的一个缩影，根子还是在大门那里。
【在 i***h 的大作中提到】
: 一厢情愿的说法
: 他这个目录和狗搜索还差好多啊
:
: that
: Google
: you
: internet,
: working
non-trivial task;"
但是在那个时代，微软整体上是在Internet的步伐上落后，甚至有意打压拖延的，即使
简单的Scale的想法也没人愿意去予以实现。这不仅是一个小的政治性的斗争的问题，
最多算是当时微软对于互联网开发缺乏预见性的一个缩影，根子还是在大门那里。
【在 i***h 的大作中提到】
: 一厢情愿的说法
: 他这个目录和狗搜索还差好多啊
:
: that
: you
: internet,
: working
p*t
11 楼
不能这么说，你只看Ohio的样本当然不够。
【在 z****r 的大作中提到】
: 非常感谢！faculty的工资看来确实比工业界的工资要低阿。
【在 z****r 的大作中提到】
: 非常感谢！faculty的工资看来确实比工业界的工资要低阿。
l*n
12 楼
微软那时候这么成功，要完全转型是不可能的。不过微软确实有机会压倒google，技术
肯定不是问题，主要是管理层的vision，那么多进去养老的烙印搞屁呀
【在 l******s 的大作中提到】
: He thought "Of course, scaling this thing to the whole internet would be a
: non-trivial task;"
: 但是在那个时代，微软整体上是在Internet的步伐上落后，甚至有意打压拖延的，即使
: 简单的Scale的想法也没人愿意去予以实现。这不仅是一个小的政治性的斗争的问题，
: 最多算是当时微软对于互联网开发缺乏预见性的一个缩影，根子还是在大门那里。
肯定不是问题，主要是管理层的vision，那么多进去养老的烙印搞屁呀
【在 l******s 的大作中提到】
: He thought "Of course, scaling this thing to the whole internet would be a
: non-trivial task;"
: 但是在那个时代，微软整体上是在Internet的步伐上落后，甚至有意打压拖延的，即使
: 简单的Scale的想法也没人愿意去予以实现。这不仅是一个小的政治性的斗争的问题，
: 最多算是当时微软对于互联网开发缺乏预见性的一个缩影，根子还是在大门那里。
l*n
13 楼
这和 bing 有神马关系？
ET
14 楼
based on the first immediate comment of this guy's answer, what he had built
wouldn't have built a search engine.
that
Google
you
internet,
working
【在 A********i 的大作中提到】
: well, I read those comments before I started this post, you may notice that
: there are also several other answers under the same question
: "Why aren't the engineers at Microsoft able to make Bing better than Google
: even though the employees at both companies are equally smart?"
: , those questions also provide some technical reasons like the comments you
: mentioned. The point in this guy's answer is that Microsoft had a big
: advantage of developing a search engine in the early stages of the internet,
: and as mentioned in this guy's answer, Microsoft had already have a working
: search engine on windows systems, but because of politics, this search
: engine never got a chance to go further.
wouldn't have built a search engine.
that
you
internet,
working
【在 A********i 的大作中提到】
: well, I read those comments before I started this post, you may notice that
: there are also several other answers under the same question
: "Why aren't the engineers at Microsoft able to make Bing better than Google
: even though the employees at both companies are equally smart?"
: , those questions also provide some technical reasons like the comments you
: mentioned. The point in this guy's answer is that Microsoft had a big
: advantage of developing a search engine in the early stages of the internet,
: and as mentioned in this guy's answer, Microsoft had already have a working
: search engine on windows systems, but because of politics, this search
: engine never got a chance to go further.
c*e
15 楼
90年代的搜索和google后的搜索完全不同。google是靠的自己建立的index和算法来做
的。9０年代的搜索，很多靠搜索html 網頁里的meta tag
Bing
equally-
【在 A********i 的大作中提到】
: 直接说是因为politics, 写得可以看出来带点情绪， 可能当时确实被politics阴得不
: 轻。
: 链接：
: https://www.quora.com/Why-arent-the-engineers-at-Microsoft-able-to-make-Bing
: -better-than-Google-even-though-the-employees-at-both-companies-are-equally-
: smart/answer/Bartosz-Milewski?share=1
: 正文：
: I worked at Microsoft in the early 90s as the project lead for the content
: index, so I know exactly why Microsoft failed in this area. The reason was
: company politics!
的。9０年代的搜索，很多靠搜索html 網頁里的meta tag
Bing
equally-
【在 A********i 的大作中提到】
: 直接说是因为politics, 写得可以看出来带点情绪， 可能当时确实被politics阴得不
: 轻。
: 链接：
: https://www.quora.com/Why-arent-the-engineers-at-Microsoft-able-to-make-Bing
: -better-than-Google-even-though-the-employees-at-both-companies-are-equally-
: smart/answer/Bartosz-Milewski?share=1
: 正文：
: I worked at Microsoft in the early 90s as the project lead for the content
: index, so I know exactly why Microsoft failed in this area. The reason was
: company politics!
相关阅读
看了大家对blockchain讨论，想起了gitAWS上做Deep learning large scale design对人的训练就像训练神经网络CSS坑真多怎么找到hash和原文的关系？美帝真的没落了d3值不值得学？如何train一个把签名（手写）与印刷字体区别的model？没有比Visual Studio更好的IDE了吧?网页上js计算的变量值现在的WINDOWS 什么还有什么应用 LINUX上面没有的？ (转载)UW的那个ML课程如何？怎么对付老邢的验证码？seq-to-seq model的参数share吗？cloud vs AI 哪个方向更适合职业发展kafka vs fluentd下一波世界革命，是小札等科技巨头忽悠P民推翻政客 (转载)跨平台开发比较最近关于DL争论挖得太慢了 (转载)