10 ms·

A Decade at Google

202 points by dwenzek 11y ago

jayvanguard 11y ago
Some sound advice. Having worked with academics before, the ones who were most successful were able to switch gears between individual and team goals as he describes.
michielr 11y ago
Down for me, Google cache link: http://webcache.googleusercontent.com/search?q=cache:WpRJJJ-4DwwJ:wp.sigmod.org/%3Fp%3D1851+&cd=1&hl=nl&ct=clnk&gl=be http://webcache.googleusercontent.com/search?q=cache:WpRJJJ-...
- chubot 11y ago
  Down for me, but is it just me or is Google's cache kinda broken? When I click the cache link I get a blank page, and then "waiting for wp.sigmod.org..." at the bottom, which hangs forever. I guess it hits the live site if it doesn't have it in the cache?
  - m_eiman 11y ago
    You can get the text-only cached version here: http://webcache.googleusercontent.com/search?q=cache:WpRJJJ-4DwwJ:wp.sigmod.org/%3Fp%3D1851&hl=nl&gl=be&strip=1&vwsrc=0 http://webcache.googleusercontent.com/search?q=cache:WpRJJJ-... The normal cached version still tries to download images and whatnot from the original server.
  - infinotize 11y ago
    Cache page doesn't hang forever for me, had to give it about 15 seconds; seems like it's trying to contact the page for images and eventually gives up and shows you the cache. Clicking "Text-only version" made it load instantly for me (there's not anything of substance but the text anyway).
mjlangiii 11y ago
Good tip for balancing work and life - just find other interests that pull you away from work.
chrisseaton 11y ago
This guy says that he's in 'Google Research' and talks about job titles like 'junior researcher'. Is there or is there not an actual Google Research organisation, similar to Oracle Labs or Microsoft Research? With people doing full-time research and publishing papers? I was under the impression that their policy was 'you can do research anywhere in Google' and that there was no separate staffed labs doing more academic research. You may get research in the V8 team, or the systems team or whatever, but they were doing research as part of product teams. At least that's what I was told when I applied.
- jlarocco 11y ago
  What a strange comment. It almost sounds like you're accusing him of lying. Anyway, it's pretty easy to verify: http://research.google.com/ http://research.google.com/ And the guy's profile: http://research.google.com/pubs/author1112.html http://research.google.com/pubs/author1112.html Edit: Don't know why I'm getting voted down. The OP was questioning whether there's a research group at Google, in the comment section for an article by somebody who's worked in that department for ten years. If that's not a definitive answer to the question, I don't really know what the OP is expecting.
  - chrisseaton 11y ago
    Google has previously made it look like there is no Google Research "even in areas where there is a much higher proportion of research to engineering, the “Research Team” we have established is not as formally separate from engineering activities as those in other organizations" so I thought it wasn't a separate group you could be on the staff of.
  - BinaryIdiot 11y ago
    > Edit: Don't know why I'm getting voted down. The OP was questioning whether there's a research group at Google, in the comment section for an article by somebody who's worked in that department for ten years. If that's not a definitive answer to the question, I don't really know what the OP is expecting. It sounded like an innocent question to me and it was one I also had. While there is a research site it still wasn't entirely obvious to me how it worked. You're probably getting downvoted for the lying bit and being condescending.
scast 11y ago
Alon Halevy is pretty big in the field of data integration. Some of my undergrad project was about improving certain aspects on his work on LAV integration. You might also find his work by Alon Levy instead of Halevy. If you want a theoretical view of the field, checkout Maurizio Lenzerini's work. For a perspective/survey check this out: http://www.dis.uniroma1.it/~degiacom/didattica/semingsoft/materiale/3-data-integration/references/Lenzerini-survey-pods-02.pdf http://www.dis.uniroma1.it/~degiacom/didattica/semingsoft/ma...
emilsedgh 11y ago
Just wanted to highlight a very important advice on the article: You do not reach [work life] balance by reducing work. You reach balance by finding a passion that draws you out of work.
- melling 11y ago
  How does that play out when you're a parent with a spouse and 4 children?
  - softawre 11y ago
    Time machine? Or find something you like to do with your kids.
  - bluedino 11y ago
    I have heard once you have children, you only have time for one hobby. True?
    - JoeAltmaier 11y ago
      Not for me. I had 3 children; also did Scouts (involving the children), wrote (1 hr every morning before everyone else woke), moderated a game club (every Thursday 4:30), visit a shooting range. Once the children are older, it accelerated of course.
      - nkassis 11y ago
        I assume you had to have good time management habits to pull it off?
        JoeAltmaier 11y ago
        Scouts was always a do-everything-at-the-last-minute thing. You just have to be present for the kids, that's the most important thing. Regularly scheduled hobbies are the easiest to pull off. If you just try to find a spare hour, you never will.
sabujp 11y ago
> (or if you cannot code, get involved in code reviews). This is the worst advice ever, esp. if you don't know how to code and don't have code readability for the language being used in the code review. Learn from code reviews, see what other people are saying, but don't comment on a fix unless you found something glaringly obvious.
- greggyb 11y ago
  I interpreted this section as saying something more along the lines of "if your job prevents you from engaging in coding as part of your day-to-day activities, but you otherwise know how to code, then maintain involvement with coding by participating in code reviews." It seems that his intended audience is expected to know how to code.
- jsolson 11y ago
  I'm a C++ developer at Google; I've been writing software "professionally"[0] for about 16 years, including a fair chunk of C, but no C++ until I reached Google. So, arguably I started in a position where commenting on code reviews was a bad idea. Nonetheless, I commented on LOTS of code reviews when I joined, but almost always with questions asking (usually) 'why is this done this way?' or (rarely) 'could this by done this other way?'. It helps that our code review tool lets you mark comments as 'no action required'. This allows the first form to be a genuine question without judgement. I would usually learn something, and sometimes the original author would revisit a decision and improve the code. I reserve the second form for when I'm pretty sure I think what I'm suggesting is an improvement. In my experience this allowed me to provide feedback without perceived ego or offense[1], but also accelerated the rate at which I came up to speed on both the language and the details of the project I'd joined. Despite being a few years in now and fairly familiar with our project, I try (although sometimes fail) to maintain this style today. [0]: My father got me my first development job as a subcontractor building terminal emulation software for the Pennsylvania Turnpike Commission at 16. I'd learned C a few years beforehand on QNX. A number of other jobs from the same folks followed over the years while I finished up high school and undergrad. [1]: Usually, anyway. There are some people who take ANY comment, question, or critique as an accusation. Thankfully I've encountered only one of these people at Google.
hrshtr 11y ago
Congrats for completing a decade and for your achievements !!
amelius 11y ago
> We then could inspect the query stream to estimate how many queries might be well served by HTML tables. Hmm.. humans are inspecting the query stream?
- hebdo 11y ago
  Obviously not, except for debugging. Scoring logic is most likely where the "secret sauce" of the thing they did is hidden.
- asuffield 11y ago
  Yes. Click here and you'll be one of them: https://www.google.com/trends/hottrends/visualize?nrow=5&ncol=5 https://www.google.com/trends/hottrends/visualize?nrow=5&nco... There's lots more data over here: http://www.google.com/trends http://www.google.com/trends (I don't believe the author was referring to this, I think he meant something far less interesting.)
geebee 11y ago
I found the part about coding interesting.. "My coding activities were always productive. They either launched a new project or a major change in direction of a project. They also enabled me to have discussions with my team members at a completely different level of detail (to everyone’s enjoyment)." This can be great. It sounds like it went well. However, I have personally seen a downside, some potential risks, when the boss "keeps coding". Here's why - if the boss is coding but staying off the critical path, this means that the often detailed, difficult implementation work will be delegated to developers who report to the boss. The old adage that genius is 1% inspiration, 99% perspiration? Well, the critical path is the perspiration part. Deadlines, estimates, glaring errors that shouldn't have happened and need to be fixed late at night or over the weekend… these are all part of the critical path. While coding can keep a boss closer to a technical team, it can also create a dangerous illusion, that the boss is still "technical" but insulated from the truly difficult aspects of a technical role. There is a big, big difference between coding on your own research project and coding under deadline pressure for a system that needs to work. This is only a risk, not a certainty, as a good manager will recognize this. Perhaps a greater risk, though, is that if the boss gets to play around with new technology and launch projects for the rest of the team to finish off, there's a risk that the boss is pretty much eating the frosting of the team's dessert but making everyone else eat their vegetables, so to speak. If you're a manager who codes, keep in mind that autonomy and innovation may be key to the motivation and job satisfaction of the people working for you. This can lead to a very damaging management anti-pattern, where a manager who "keeps coding" essentially uses managerial authority to intercept and filter projects, make the key technical decisions, do the interesting bits that would lead to interesting conference talks, and then delegate the work of actually finishing the product to the team. This may deprive senior level developers the autonomy that they really were supposed to have on projects. Once again, I want to emphasize that this isn't necessarily what happened here, in fact, it sounds like it wasn't. But the passage made me a bit nervous, since I consider this to be something of a management anti-pattern.
- gavazzy 11y ago
  A manager who doesn't continue to code or work on the day-to-day actions may fail to keep up with necessary technical knowledge, effectively reducing their ability to manage.

202 points by dwenzek 11y ago

Some sound advice. Having worked with academics before, the ones who were most successful were able to switch gears between individual and team goals as he describes.

michielr 11y ago

Down for me, Google cache link: http://webcache.googleusercontent.com/search?q=cache:WpRJJJ-4DwwJ:wp.sigmod.org/%3Fp%3D1851+&cd=1&hl=nl&ct=clnk&gl=be http://webcache.googleusercontent.com/search?q=cache:WpRJJJ-...

chubot 11y ago
Down for me, but is it just me or is Google's cache kinda broken? When I click the cache link I get a blank page, and then "waiting for wp.sigmod.org..." at the bottom, which hangs forever. I guess it hits the live site if it doesn't have it in the cache?
- m_eiman 11y ago
  You can get the text-only cached version here: http://webcache.googleusercontent.com/search?q=cache:WpRJJJ-4DwwJ:wp.sigmod.org/%3Fp%3D1851&hl=nl&gl=be&strip=1&vwsrc=0 http://webcache.googleusercontent.com/search?q=cache:WpRJJJ-... The normal cached version still tries to download images and whatnot from the original server.
- infinotize 11y ago
  Cache page doesn't hang forever for me, had to give it about 15 seconds; seems like it's trying to contact the page for images and eventually gives up and shows you the cache. Clicking "Text-only version" made it load instantly for me (there's not anything of substance but the text anyway).

mjlangiii 11y ago

Good tip for balancing work and life - just find other interests that pull you away from work.

chrisseaton 11y ago

This guy says that he's in 'Google Research' and talks about job titles like 'junior researcher'. Is there or is there not an actual Google Research organisation, similar to Oracle Labs or Microsoft Research? With people doing full-time research and publishing papers? I was under the impression that their policy was 'you can do research anywhere in Google' and that there was no separate staffed labs doing more academic research. You may get research in the V8 team, or the systems team or whatever, but they were doing research as part of product teams. At least that's what I was told when I applied.

jlarocco 11y ago
What a strange comment. It almost sounds like you're accusing him of lying. Anyway, it's pretty easy to verify: http://research.google.com/ http://research.google.com/ And the guy's profile: http://research.google.com/pubs/author1112.html http://research.google.com/pubs/author1112.html Edit: Don't know why I'm getting voted down. The OP was questioning whether there's a research group at Google, in the comment section for an article by somebody who's worked in that department for ten years. If that's not a definitive answer to the question, I don't really know what the OP is expecting.
- chrisseaton 11y ago
  Google has previously made it look like there is no Google Research "even in areas where there is a much higher proportion of research to engineering, the “Research Team” we have established is not as formally separate from engineering activities as those in other organizations" so I thought it wasn't a separate group you could be on the staff of.
- BinaryIdiot 11y ago
  > Edit: Don't know why I'm getting voted down. The OP was questioning whether there's a research group at Google, in the comment section for an article by somebody who's worked in that department for ten years. If that's not a definitive answer to the question, I don't really know what the OP is expecting. It sounded like an innocent question to me and it was one I also had. While there is a research site it still wasn't entirely obvious to me how it worked. You're probably getting downvoted for the lying bit and being condescending.

scast 11y ago

Alon Halevy is pretty big in the field of data integration. Some of my undergrad project was about improving certain aspects on his work on LAV integration. You might also find his work by Alon Levy instead of Halevy. If you want a theoretical view of the field, checkout Maurizio Lenzerini's work. For a perspective/survey check this out: http://www.dis.uniroma1.it/~degiacom/didattica/semingsoft/materiale/3-data-integration/references/Lenzerini-survey-pods-02.pdf http://www.dis.uniroma1.it/~degiacom/didattica/semingsoft/ma...

emilsedgh 11y ago

Just wanted to highlight a very important advice on the article: You do not reach [work life] balance by reducing work. You reach balance by finding a passion that draws you out of work.

melling 11y ago
How does that play out when you're a parent with a spouse and 4 children?
- softawre 11y ago
  Time machine? Or find something you like to do with your kids.
- bluedino 11y ago
  I have heard once you have children, you only have time for one hobby. True?
  - JoeAltmaier 11y ago
    Not for me. I had 3 children; also did Scouts (involving the children), wrote (1 hr every morning before everyone else woke), moderated a game club (every Thursday 4:30), visit a shooting range. Once the children are older, it accelerated of course.
    - nkassis 11y ago
      I assume you had to have good time management habits to pull it off?
      - JoeAltmaier 11y ago
        Scouts was always a do-everything-at-the-last-minute thing. You just have to be present for the kids, that's the most important thing. Regularly scheduled hobbies are the easiest to pull off. If you just try to find a spare hour, you never will.

sabujp 11y ago

> (or if you cannot code, get involved in code reviews). This is the worst advice ever, esp. if you don't know how to code and don't have code readability for the language being used in the code review. Learn from code reviews, see what other people are saying, but don't comment on a fix unless you found something glaringly obvious.

greggyb 11y ago
I interpreted this section as saying something more along the lines of "if your job prevents you from engaging in coding as part of your day-to-day activities, but you otherwise know how to code, then maintain involvement with coding by participating in code reviews." It seems that his intended audience is expected to know how to code.
jsolson 11y ago
I'm a C++ developer at Google; I've been writing software "professionally"[0] for about 16 years, including a fair chunk of C, but no C++ until I reached Google. So, arguably I started in a position where commenting on code reviews was a bad idea. Nonetheless, I commented on LOTS of code reviews when I joined, but almost always with questions asking (usually) 'why is this done this way?' or (rarely) 'could this by done this other way?'. It helps that our code review tool lets you mark comments as 'no action required'. This allows the first form to be a genuine question without judgement. I would usually learn something, and sometimes the original author would revisit a decision and improve the code. I reserve the second form for when I'm pretty sure I think what I'm suggesting is an improvement. In my experience this allowed me to provide feedback without perceived ego or offense[1], but also accelerated the rate at which I came up to speed on both the language and the details of the project I'd joined. Despite being a few years in now and fairly familiar with our project, I try (although sometimes fail) to maintain this style today. [0]: My father got me my first development job as a subcontractor building terminal emulation software for the Pennsylvania Turnpike Commission at 16. I'd learned C a few years beforehand on QNX. A number of other jobs from the same folks followed over the years while I finished up high school and undergrad. [1]: Usually, anyway. There are some people who take ANY comment, question, or critique as an accusation. Thankfully I've encountered only one of these people at Google.

hrshtr 11y ago

Congrats for completing a decade and for your achievements !!

amelius 11y ago

> We then could inspect the query stream to estimate how many queries might be well served by HTML tables. Hmm.. humans are inspecting the query stream?

hebdo 11y ago
Obviously not, except for debugging. Scoring logic is most likely where the "secret sauce" of the thing they did is hidden.
asuffield 11y ago
Yes. Click here and you'll be one of them: https://www.google.com/trends/hottrends/visualize?nrow=5&ncol=5 https://www.google.com/trends/hottrends/visualize?nrow=5&nco... There's lots more data over here: http://www.google.com/trends http://www.google.com/trends (I don't believe the author was referring to this, I think he meant something far less interesting.)

geebee 11y ago

I found the part about coding interesting.. "My coding activities were always productive. They either launched a new project or a major change in direction of a project. They also enabled me to have discussions with my team members at a completely different level of detail (to everyone’s enjoyment)." This can be great. It sounds like it went well. However, I have personally seen a downside, some potential risks, when the boss "keeps coding". Here's why - if the boss is coding but staying off the critical path, this means that the often detailed, difficult implementation work will be delegated to developers who report to the boss. The old adage that genius is 1% inspiration, 99% perspiration? Well, the critical path is the perspiration part. Deadlines, estimates, glaring errors that shouldn't have happened and need to be fixed late at night or over the weekend… these are all part of the critical path. While coding can keep a boss closer to a technical team, it can also create a dangerous illusion, that the boss is still "technical" but insulated from the truly difficult aspects of a technical role. There is a big, big difference between coding on your own research project and coding under deadline pressure for a system that needs to work. This is only a risk, not a certainty, as a good manager will recognize this. Perhaps a greater risk, though, is that if the boss gets to play around with new technology and launch projects for the rest of the team to finish off, there's a risk that the boss is pretty much eating the frosting of the team's dessert but making everyone else eat their vegetables, so to speak. If you're a manager who codes, keep in mind that autonomy and innovation may be key to the motivation and job satisfaction of the people working for you. This can lead to a very damaging management anti-pattern, where a manager who "keeps coding" essentially uses managerial authority to intercept and filter projects, make the key technical decisions, do the interesting bits that would lead to interesting conference talks, and then delegate the work of actually finishing the product to the team. This may deprive senior level developers the autonomy that they really were supposed to have on projects. Once again, I want to emphasize that this isn't necessarily what happened here, in fact, it sounds like it wasn't. But the passage made me a bit nervous, since I consider this to be something of a management anti-pattern.

gavazzy 11y ago
A manager who doesn't continue to code or work on the day-to-day actions may fail to keep up with necessary technical knowledge, effectively reducing their ability to manage.