5

This is a more basic question than, and not a duplicate of, this question. That question was answered, basically, with "become familiar with the literature in that field" and read that literature. So that question is answered, but leaves me with this more basic one.

I'm a senior software engineer albeit with only a BS in CS. I'm not that in touch with new ideas in such as red-black trees (not new now but newer than my undergrad degree), the latest ideas on to optimize hash tables, and the like, except as they pick up use in commonly used libraries such as Boost, or otherwise spread in the industry. My questions are:

  1. I'd like to be more of an early adopter. How do I even find such journals or other publications? Note I'm with a financial firm in Japan and I doubt it has any journal subscriptions. I'm not sure how to see if any local universities make their libraries open to non-affiliated independent researchers.

  2. I've developed a new type of memory pool that can maximize locality of reference of allocations. That should allow an application to substantially improve performance. I'd like to publish it, if 1) it's not "too practical" for a research journal and 2) if a survey of the literature shows no-one's already invented it. OTOH, I don't know how to do that survey.

Alternately, is simply trying to put something like this in Boost or GNU libraries or on github the most reasonable approach?

  • Is the memory pool physical or an adaptation of the current hardware? Can you patent it? – Poidah Oct 7 '19 at 3:27
  • It's portable software; runs on any CPU. It was designed to achieve benefits on Intel caches (more app data per cache line, reduce 4k aliasing, can eliminate false sharing if that's a problem, reduces working set, etc. etc.) but probably has similar benefits on most other processors. I've never looked into software patenting but it could be patentable. I've been wondering about the utility of that but I suppose that's a separate question I should look into outside the scope of this question. – Swiss Frank Oct 7 '19 at 3:37
  • 1
    Perhaps start following academics who do work you are interested in on twitter etc, we inevitably post links to our work. That's nice and free. It's my impression that hot new computer science research comes out in conferences, not journals. You can attend those conferences as a non academic. Most of their proceedings are open access. – GrotesqueSI Oct 7 '19 at 7:26
  • Thanks GrotesqueSI ... it sort of leaves me wondering who these academics would be... I suppose I could just find recently-developed algos on Wiki to get an initial set of authors to follow. Still, I think the days where single researchers invent lots of things in Computer Science may be far behind us. In the 50s-70s single guys invented whole worlds of software, but the modern day has me wondering if any single person ever even gets to three notable inventions. – Swiss Frank Oct 8 '19 at 0:03
  • Before you consider trying to publish a paper on work you’ve done, consult your employment contract to make sure you actually own the rights to your invention; many people employed in IT positions have contracts that state that the employer owns all inventions produced by their employees during the time period of their employment, even if they’re produced outside of working hours. If they do, then you’ll definitely need to ask their permission before publishing. – nick012000 Oct 8 '19 at 4:11
1

Maybe look for review journals on the topic. In my area there are several such journals. Researchers often do a "state of the topic" paper on something they are working on. You can scan a few 100 review topic titles. Should take you three or four hours. Then find three or four articles that you can zero in on. You should be able to "surface" read these articles in a couple days. Your university librarian should be able to help you find appropriate review journals. Once you have those papers you can also read the citations they tell you are most important.

| improve this answer | |
  • Fair enough. What'd be the first step in even getting access to a university library? Here in Japan I'm not sure non-affiliated persons are allowed on campus, or allowed in libraries. I suppose I could just get a local university switchboard and call and ask... – Swiss Frank Oct 7 '19 at 23:59
  • 1
    In the US our university libraries have websites. Each website includes staff profiles. The staff typically include subject librarians who are specialized in particular departments/topics. – Dawn Oct 8 '19 at 3:20
  • 1
    Most places the university libraries have some degree of sharing with public libraries. According to this web site there are over 3000 public libraries in Japan. jla.or.jp/portals/0/html/libraries-e.html Go to a public library and ask for assistance. It may mean you have to wait a week or two to get a given article. Also, many journals have on-line options. You might need to pay to get the articles, and sometimes that's prohibitively expensive. Maybe you can get single articles. – puppetsock Oct 8 '19 at 13:45

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.