The Ur-Quan Masters Discussion Forum

(https://i.imgur.com/RxXry6s.png)

Who's paying for the databases, and for editing at the journals?

Who is paying for Linux? I guess the databases can be open source, and lots of the work with incorporating data from journals into databases can be done with machine learning algorithms today.

Sure, the editors at the journals should be payed, but I doubt they receive most of the money here anyhow, and what about the researchers that wrote the articles in the first place? They aren't payed at all, and in many cases the editors barely need to make any corrections, and in many journals they don't necessarily see errors. At least I often find errors in many journals.

BTW, a shared license for 1 user at a university, cost approximately 12 000 USD/Year just for SciFinder, and the universities usually need access to multiple shared users in multiple similar databases with similar prices. Regarding access to Elsevier:


https://en.wikipedia.org/wiki/Elsevier#Boycotts (https://en.wikipedia.org/wiki/Elsevier#Boycotts)

Finland is a reasonably rich western European country, so if they are struggling with this there, imagine what the situation must be like in poor countries.

Do you think it is reasonable to have a system which makes it more or less impossible for people to do cutting edge research in poor countries?

I am also adding this:

(https://i.imgur.com/XeSCLhE.png)

I have now published it at: http://www.archania.org/governance/#Linux_and_open_access_to_scientific_journals_and_databases (http://www.archania.org/governance/#Linux_and_open_access_to_scientific_journals_and_databases)

Need a new machine, will likely test Linux on the old one, once the data has been migrated. Thanks for the advice ("Cinnamon").

Plenty of universities tried Linux.
The crux is and remains that the hardware systems are so diverse, that testing all hardware where the stuff needs to run is impossible for the university administrators.

On top of that, many institutes at universities are very independent from the central procurement, and do therefore not follow the same guidelines.

And many software needed to run the experiments are often provided Windows-only.

I know I ran my studies on three machines, 2 windows (one running windows only software from the attached test machine supplier), one Linux.
We had most issues with the Linux machine.
All kind of upgrade issues,.... New hardware -> not functioning or somehow causing problems with the older necessary hardware parts, new hardware sometimes required updates for the kernel, which older hardware elements could not be upgraded to (old testing equipment, no new drivers available).
With windows, the core remained. The whole system just remained functional. No matter what we did install/uninstall/changed in the settings/...

That was 15 years ago.
Linux got much better in that regard, but back then our institute decided to remain windows only for the time being. (So we got new hardware, and migrated the lab machine equipment to the new wondows computer.)
A new try today would likely work out much better than back then.
But I fear the basic thought which caused us to elect windows remains.
If MS ever changes their licensing fees, so that you have to pay annually or similar, then the equation will change (too much money spent on software licenses anyway, money is better spent on the equipment), until then, we have a warm heart for Linux, but Windows remains the cheaper option for many institutes.


Regarding publications: in the Netherlands and in Germany, Elsevier and the like now give huge discounts to the universities still publishing via their scientific magazines.
The universities found it unfair, that they provided the studies to be published, some publisher made money, and the universities had to subscribe and pay to get their own studies...

The bigger scientific publications still pay to double-check the content of the publications, and this costs a lot of money.
"Free publication servers" are taking a hold too (especially for public-money funded researches), but the publications there remain "unconfirmed", and thus are not as desireable for publications and citing.

This is why I changed from Linux Mint to Manjaro/Arch Linux a while ago. With my Manjaro-Cinnamon combination, I get updates immediately when a new kernel is released, and it is super easy to upgrade:

(https://i.imgur.com/G8qKoge.png)

(https://i.imgur.com/xSL9AdE.png)

The newest kernels are often compatible with most of the newest hardware. They are at 4.15.11 now, so I am a little bit behind, but not a lot: www.kernel.org (http://www.kernel.org)

I have been using Linux for at least 15 years now, and I would say that it is much better now than back then. I have also experimented with different distributions and desktop environments recently, and I definitely like the Manjaro-Cinnamon combination best.


Maybe they give discounts to huge universities in rich countries, which are in a strong position to bargain with them. I doubt they give a lot of discount to small universities, and universities in poor countries.

But then, the universities in the big countries are the ones who publish the most in the scientific publications.
That was the leverage to enforce a deduction in subscription fees.
(Otherwise the German universities would've left the "old" system and would publish via their own online publication system.
I find the latter still a good idea, and I hope they'll implement it anyway sometime soon.)

Sure, but how are universities with less bargain power ever going to get better, if they don't even have access to relevant scientific publications and databases?

(https://i.imgur.com/ljmMCL5.png)

https://www.archania.org/research/#Academic_research_articles_should_be_rated (https://www.archania.org/research/#Academic_research_articles_should_be_rated)

For academic articles that aren't about chemical reactions, we can just put the "abstract" instead. I am going to put all the references from my website into a database where people can rate them, and write comments. I will also put lots of chemical reactions into the database, and make it open so that people can add new articles or reactions.

Okay.... here is the beta version of my database with rated chemical reactions:

https://www.archania.org/research/reactions/?name=&year=&type=&sortby=name (https://www.archania.org/research/reactions/?name=&year=&type=&sortby=name)

I am thinking that chemists tend to be reasonably smart people, so I want chemists to come to my website. And if I put a lot of chemical reactions into the database, chemists from all over the world might be lured to my website when they search for chemical reactions in google. Seems like it could be a quite good tractor beam generator.

If I made a similar database with mathematical theorems, I might be able to lure mathematicians from all over the world to my website....

I have now added 80 chemical reactions, and you can write comments and rate anyone of them anonymously without any account.

The reactions can also be sorted according to how they are rated: https://www.archania.org/research/reactions/ (https://www.archania.org/research/reactions/)

Google has started indexing the reactions on my webpage now: https://www.google.com/search?q=site:archania.org/research/reactions&source=lnms&tbm=isch (https://www.google.com/search?q=site:archania.org/research/reactions&source=lnms&tbm=isch)

I would like to make it so that people can draw chemical motifs, and search for reactions fitting the motifs. Unfortunately, this seems to be a bit beyond my skills as a programmer. :(

They have it like that in the 10 000 USD per year per user database called SciFinder. If my database also got something like that, it could become a serious free competitor to super expensive databases like SciFinder and Reaxys.

Universities can also have their own public databases, just look at how popular Cornell's ArXiv is, which you can also find here https://arxiv.org/

Yes. I have written a bit more about why universities should publish in open access journals today:

(https://i.imgur.com/Ayddrnj.png)

https://www.archania.org/research/#Universities_should_use_Linux_and_publish_in_open_access_journals (https://www.archania.org/research/#Universities_should_use_Linux_and_publish_in_open_access_journals)

Also... if you think open access journals don't provide as much quality control as expensive closed access journals, I have also written about how all researchers should be allowed to rate scientific articles, similarly to how you might rate a movie on imdb.com.

Anyhow, I see that the problem with closed access publishing is particularly prevalent with chemistry:

(https://dfzljdn9uc3pi.cloudfront.net/2018/4375/1/fig-4-2x.jpg)

https://peerj.com/articles/4375/ (https://peerj.com/articles/4375/)

I bought a separate domain for the database with rated chemical reactions. I might hire people to work for me to make this database better.

https://www.ratedreactions.com/ (https://www.ratedreactions.com/)

Now, this is only going to be used by big companies, and rich universities, since it is so expensive. If we can make an open source version of something like this, it might benefit all of humanity to a much higher degree.

Look at these guys:


https://motherboard.vice.com/en_us/article/43pngb/how-to-make-your-own-medicine-four-thieves-vinegar-collective (https://motherboard.vice.com/en_us/article/43pngb/how-to-make-your-own-medicine-four-thieves-vinegar-collective)

Two women are starting to work with adding reactions to https://www.ratedreactions.com (https://www.ratedreactions.com) tomorrow, so hopefully, I will soon have much more reactions there.

I am trying to figure out how to make canonical SMILES. There is an article about it here, http://ctr.wikia.com/wiki/Convert_a_SMILES_string_to_canonical_SMILES (http://ctr.wikia.com/wiki/Convert_a_SMILES_string_to_canonical_SMILES), but I am programming with PHP/Javascript/HTML, and lots of those algorithms seem to be written in C which I don't think I can execute from my web server. I think it should be possible to convert a SMILES string to canonical SMILES in PHP. I only need to use this algorithm before I save SMILES into the database, and after people have drawn reaction schemes. So it doesn't necessarily need it to go as fast as C code.

(https://i.imgur.com/BjWsID1.png)

Any help would be greatly appreciated.

how large are these programs? I suspect they could be small, in which case converting from C to php could take a modest amount of time for anyone familiar with both languages.

I also think it would be easy for someone that is familiar with many different programming languages.

I am not educated as a software engineer, and I am not very familiar with C, Python, or Java.  I am most familiar with PHP, Javascript, HTML, and LaTeX.

Here is a C-algorithm for making canonical SMILES:

https://github.com/epam/Indigo/blob/master/utils/indigo-cano/main.c (https://github.com/epam/Indigo/blob/master/utils/indigo-cano/main.c)

that file is not so big, but the include of indigo.h makes it big enough that this would not be lightly done.

I see, but all the functions in indigo.h are probably not necessarily used, so maybe only some parts of indigo.h needs to be converted.

Anyhow, I guess I could hire some freelancer to do it for me from one of these sites:  https://fitsmallbusiness.com/best-freelance-websites-small-business/ (https://fitsmallbusiness.com/best-freelance-websites-small-business/)