5 ms·

You can get a Phd by doing stuff like this?? I do that kind of thing for fun after work on my hobby projects, and at work I've spent weeks speeding up a single

by brabel 2y ago

You can get a Phd by doing stuff like this?? I do that kind of thing for fun after work on my hobby projects, and at work I've spent weeks speeding up a single HTTP request path.

FrostKiwi 2y ago
Highly specialized topic of computer science, in a tough challenge with no easy solutions and new discoveries documented for anyone to follow in the future. I would be sad, if you couldn't get a Phd for stuff like this.
kragen 2y ago
theoretically your result has to be (1) novel and (2) published with such a good explanation that people understand both the result and what's novel about it, as well as how to reproduce it, but yes. the difference between tinkering and science is documentation, and it needs to be public and conscious of the existing discourse to be scholarship
- gwervc 2y ago
  The biggest difference is gate-keeping by peer review, which is quite random. Reproductibility is a meme in most fields.
  - jvanderbot 2y ago
    Even in random noise, with enough samples the signal is measurable. Thus - publish or perish.
  - kragen 2y ago
    i knew someone was going to reply by pointing out that academia often fails to live up to the standards of scholarship and, where applicable, science. but that's sort of missing the point; i was describing the standards, not how the existing institutions' performance measures up to them
- brabel 2y ago
  No offence, but what's novel in this post??? I found it pretty basic, or?
  - OJFord 2y ago
    This is a blog post, not published peer-reviewed research. You don't 'get a PhD' for a blog post, no.
  - kragen 2y ago
    not every brick in the taj mahal is beautiful
lifthrasiir 2y ago
While I don't have a PhD---I do have a master's degree in PL and have some published papers though---, yes you can do because a PhD only means that you can do a level of research required for the academia, at least in the technical sense. (PhD also suggests, but not always implies that you know enough tacit knowledge about a certain academic area.) You can even publish papers yourself if you can go through the usual academic process like peer review [1]. [1] Sadly enough, the current possibility of fake research means that even publishing to arXiv requires some support from the current researcher.
- gwervc 2y ago
  > You can even publish papers yourself It's becoming harder and harder because of APC fees :(
  - lifthrasiir 2y ago
    Indeed, that's why I mentioned arXiv.
  - nimish 2y ago
    Yeah this is kind of nuts: >For the majority of IEEE’s fully open access journals, the article processing fee will be $1,995 USD. Some exceptions apply for certain titles, see individual journal author instructions for specific details. Reviewers don't get paid, journal editorial board members usually don't get paid, and it's all online so publishing is basically free. Rent seeking at its finest!
myworkinisgood 2y ago
There are three things here: 0. A PhD is a starting degree in research, in that, it shows that you can do research. So, a lot of industry work (after you have experience for some years) can be equivalent to PhD work. 1. An obvious exaggeration but somewhat true: https://xkcd.com/664/ https://xkcd.com/664/ 2. Apart from tinkering and actually achieving a result, doing a PhD often also means being able to know why it makes sense to explore the field. So, in addition to actually achieving speedup, you have to study the background of all approaches and being aware of advantages and disadvantages of all of them.
gsck 2y ago
The only thing that differentiates having fun and science, is writing it down.
- brabel 2y ago
  Right, and I do that sometimes by writing blog posts. But as other commenter said, I would still have to: * get peer reviewed. * mention previous literature (that sucks, but I agree is needed to show an understanding of the topic). * introduce novel concepts. The barrier is high.
a1o 2y ago
If you do this by hobby, if you ever have free time and want to play with c++11, here is an interpreter that is dear to my heart that I am curious if someone can figure how to speed it up - the tricky parts are the function calls. https://github.com/adventuregamestudio/ags/blob/ags4/Engine/script/cc_instance.cpp https://github.com/adventuregamestudio/ags/blob/ags4/Engine/...

by brabel 2y ago

You can get a Phd by doing stuff like this?? I do that kind of thing for fun after work on my hobby projects, and at work I've spent weeks speeding up a single HTTP request path.

FrostKiwi 2y ago

Highly specialized topic of computer science, in a tough challenge with no easy solutions and new discoveries documented for anyone to follow in the future. I would be sad, if you couldn't get a Phd for stuff like this.

kragen 2y ago

theoretically your result has to be (1) novel and (2) published with such a good explanation that people understand both the result and what's novel about it, as well as how to reproduce it, but yes. the difference between tinkering and science is documentation, and it needs to be public and conscious of the existing discourse to be scholarship

gwervc 2y ago
The biggest difference is gate-keeping by peer review, which is quite random. Reproductibility is a meme in most fields.
- jvanderbot 2y ago
  Even in random noise, with enough samples the signal is measurable. Thus - publish or perish.
- kragen 2y ago
  i knew someone was going to reply by pointing out that academia often fails to live up to the standards of scholarship and, where applicable, science. but that's sort of missing the point; i was describing the standards, not how the existing institutions' performance measures up to them
brabel 2y ago
No offence, but what's novel in this post??? I found it pretty basic, or?
- OJFord 2y ago
  This is a blog post, not published peer-reviewed research. You don't 'get a PhD' for a blog post, no.
- kragen 2y ago
  not every brick in the taj mahal is beautiful

lifthrasiir 2y ago

While I don't have a PhD---I do have a master's degree in PL and have some published papers though---, yes you can do because a PhD only means that you can do a level of research required for the academia, at least in the technical sense. (PhD also suggests, but not always implies that you know enough tacit knowledge about a certain academic area.) You can even publish papers yourself if you can go through the usual academic process like peer review [1]. [1] Sadly enough, the current possibility of fake research means that even publishing to arXiv requires some support from the current researcher.

gwervc 2y ago
> You can even publish papers yourself It's becoming harder and harder because of APC fees :(
- lifthrasiir 2y ago
  Indeed, that's why I mentioned arXiv.
- nimish 2y ago
  Yeah this is kind of nuts: >For the majority of IEEE’s fully open access journals, the article processing fee will be $1,995 USD. Some exceptions apply for certain titles, see individual journal author instructions for specific details. Reviewers don't get paid, journal editorial board members usually don't get paid, and it's all online so publishing is basically free. Rent seeking at its finest!

myworkinisgood 2y ago

There are three things here: 0. A PhD is a starting degree in research, in that, it shows that you can do research. So, a lot of industry work (after you have experience for some years) can be equivalent to PhD work. 1. An obvious exaggeration but somewhat true: https://xkcd.com/664/ https://xkcd.com/664/ 2. Apart from tinkering and actually achieving a result, doing a PhD often also means being able to know why it makes sense to explore the field. So, in addition to actually achieving speedup, you have to study the background of all approaches and being aware of advantages and disadvantages of all of them.

gsck 2y ago

The only thing that differentiates having fun and science, is writing it down.

brabel 2y ago
Right, and I do that sometimes by writing blog posts. But as other commenter said, I would still have to: * get peer reviewed. * mention previous literature (that sucks, but I agree is needed to show an understanding of the topic). * introduce novel concepts. The barrier is high.

a1o 2y ago

If you do this by hobby, if you ever have free time and want to play with c++11, here is an interpreter that is dear to my heart that I am curious if someone can figure how to speed it up - the tricky parts are the function calls. https://github.com/adventuregamestudio/ags/blob/ags4/Engine/script/cc_instance.cpp https://github.com/adventuregamestudio/ags/blob/ags4/Engine/...