Tag Archives: Open Science

Open Science is rwx science

Preamble: this is a short text on Open Science, written a while ago,  which I now put it here. It is taken from this place at telegra.ph. The link (not the content) appeared here at the Chemlambda for the people post. I can’t find other traces, except the empty github repository “creat”,  described as “framework for research output as a living creature“.

__________________

I am a big fan of Open Science. For me, a good piece of research is one which I can Read Write eXecute.

Researchers use articles to communicate. Articles are not eXecutable. I can either Read others’ articles or Write mine. I have to trust an editor who tells me that somebody else, whom I don’t know, read the article and made a peer-review.

No. Articles are stories told by researchers about how they did the work. And since the micromanagement era, they are even less: fungible units to be used in funding applications, by the number or by the keyword.

This is so strange. I’m a mathematician and you probably know that mathematics is the most economical way to explain something clearly. Take a 10 pages research article. It contains the intensive work of many months. Now, compress the article further more by the following ridiculous algorithm: throw away everything but the first several bits. Keep only the title, the name of the journal, keywords, maybe the Abstract. That’s not science communication, that’s massive misuse of brain material.

So I’m an Open Science fan, what should I do instead of writing articles? Maybe I should push my article in public and wait after that for somebody to review it. That’s called Open Access and it’s very good for the readers. So what? the article is still only Readable or Writable, pick only one option, otherwise it’s bad practice. What about my time? It looks that I have to wait and wait for all the bosses, managers, politicians and my fellow researchers to switch to OA first.

It’s actually much easier to do Open Science, remember! something that you can Read, Write and eXecute. As an author, you don’t have to wait for the whole society to leave the old ways and to embrace the new ones. You can just push what you did: stories, programs, data, everything. Any reader can pull the content and validate it, independently. EXecute what you pushed, Read your research story and Write derivative works.

I tried this! Want to know how to build a molecular computer which is indiscernible from how we are made? Use this playground called chemlambda. It’s a made up, simple chemistry. It works like the real chemistry does, that is locally, randomly, without any externally imposed control. My bet is that chemlambda can be done in real life. Now, or in a few years.

I use everything available to turn this project into Open Science. You name it: old form articles, html and javascript articles, research blog, Github repository, Figshare data repository, Google collection [update: deleted], this 🙂

Funny animations obtained from simulations. Those simulations can be run on your computer, so you can validate my research. Here’s what chemlambda looks like.

[Here come some examples and animations. ]

 

During this project I realized that it went beyond a Read Write Execute thing. What I did was to design many interesting molecules. They work by themselves, without any external control. Each molecule is like a theorem and the chemical evolution is the proof of the theorem, done by a blind, random, stupid, universal algorithm.

Therefore my Open Science attempt was to create molecules, some of them exhibiting a metabolism, some of them alive. Maybe this is the future of Open Science. To create a living organism which embodies in its metabolism the programs and research data. It’s valid if it lives, grow, reproduces, even die. Let it cross breed with other living creatures. In time the natural selection will do marvels. Life is not different than Science. Science is not different than life.

Advertisements

Open Science: “a complete institution for the use of learners”

The quote is from 1736. You can see it on the front page of the book “The method of fluxions and infinite series” by Newton, “translated from the author’s Latin original not yet made publick” (nobody is perfect, we know now where this secrecy led in the dispute with Leibniz over the invention of the differential calculus).

newton

That should be the goal of any open science research output.

What we have at the end of 2017?

  • Sci-hub. Pros: not corporate. It does not matter where you output your article, as long as it becomes available to any learner. Cons:  only old style articles, not more. So not a full solution.
  • ArXiv. Pros: simple, proved to be reliable long term. Cons: only articles.
  • Zenodo. Pros: not corporate, lots of space for present needs. Cons: not playable.
  • Github. Pros: good for publicly and visibly share and discuss over articles and programs. Cons: corporate, not reliable in the long term.
  • Git in general. Pros: excellent tool.
  • Blockchain. Pros: excellent tool.

I have not added anything about BOAI inspired Open Access because it is something from the past. It was just a trick to delay the demise of legacy publishing style, it was done over the heads of researchers, basically a deal between publishers and academic managers, for them to be able to siphon research $  and stiffle the true open access movement.

Conclusion: at the moment there are only timid and partial proposals for open science as “a complete institution for the use of learners”. Open science is not a new idea. Open science is the natural way to do science.

There is only one way to do it: share. Let’s do it!

Transparency is superior to trust

I am fascinated by this quote. I think it’s the most beautiful quote, in it’s terseness, I’ve seen since a long time. Wish I invented it!

It is not, though, the motto of Wikileaks, it’s taken from the section on Reproducibility of this Open Science manifesto.

To me, this quote means that validation is superior to peer review.

It is also significant that the quote says nothing about the publishing aspects of Open Science. That is because, I believe, we should split publishing from the discussion about Open Science.

Publishing, scientific publishing I mean, is simply irrelevant at this point. The strong part of Open Science, the new, original idea it brings forth is validation.

Sci-Hub acted as the great leveler, as concerns scientific publication. No interested reader cares, at this point, if an article is hostage behind a paywall or if the author of the article paid money for nothing to a Gold OA publisher.

Scientific publishing is finished. You have to be realistic about this thing.

But science communication is a far greater subject of interest. And validation is one major contribution to a superior scientific method.

More experiments with Open Science

I still don’t know which format is better for Open Science. I’m long past the article format for obvious reasons. Validation is a good word and concept because you don’t have to rely absolutely on opinions of others and that’s how the world works. This is not all the story though.

I am very fortunate to be a mathematician, not a biologist or biochemist. Still I long for the good format for Open Science, even if, as a mathematician, I don’t have the problems biologists or chemists have, namely loads and loads of experimental data and empirical approaches. I do have a world of my own to experiment with, where I do have loads of data and empirical constructs. My mind, my brain are real and I could understand myself by using tools of chemists and biologists to explore the outcomes of my research. Funny right? I can look at myself from the outside.

That is why  I chose to not jump directly to make Hydrogen, but instead to treat the chemlambda  world, again, as a guinea pig for Open Science.

There are 427 well written molecules in the chemlambda library of molecules on Github. There are 385 posts in the chemlambda collection on Google+, most of them with animations from simulations of those molecules. It is a world, how big is it?

It is easy to make first a one page direct access to the chemlambda collection. It is funnier to build a phylogenetic tree of the molecules, based on their genes. That’s what I am doing now, based on a work in progress.

Each molecule can be decomposed in “genes” say, by a sequencer program. Then one can use a distance between these genes to estimate first how they cluster and later to make a phylogenetic tree.

Here is the first heatmap (using the edit distance between single occurrences of genes in molecules) of the 427 molecules.

Screenshot-22

Is a screenshot, proving that my custom programs work 🙂 (one understands more by writing some scripts than by taking tools ready made from others, at least at this stage of research).

By using the edit distance I can map the explored chemlambda molecules. In the following image the 427 molecules from the library are represented as nodes and for each pair of molecules at an edit distance at most 20 there is a link. The nodes are in a central gravitational field, each node has the same charge and the links between nodes act as springs.

Screenshot-29_map

This is a screenshot of the result, showing clusters and trees, connecting them. Not very sophisticated, but enough to give a sense of the explored territory. In the curated collection, such a map would be useful to navigate through the molecules, as well as for giving ideas about which parts are not as well explored. I have not yet made clear which parts of the map cover lambda terms, which cover quines, etc.

Moreover, I see structure! The 427 molecules are made of copies of  605 different linear “genes” (i.e. sticks with colored ends)  and 38 ring shaped ones.  (Is easy to prove that lambda terms have no rings, when turned into molecules.) There are some interesting curved features visible in the edit distance of the sticks.

screenshot-23

They don’t look random enough.

Is clear that a phylogenetic tree is in reach, then what else than connecting the G+ collection posts with the molecules used, arranged along the tree…?

Can I discover which molecules are coming from lambda terms?

Can I discover how my mind worked when building these molecules?

Which are the neglected sides, the blind places?

I hope to be able to tell by the numbers.

Which brings me to the main subject of this post: which is a good format for an Open Science piece of research?

Right now I am in between two variants, which may turn out to not be as different as they seem. An OS research vehicle could be:

  • like a viable living organism, literary
  • or like a viable world, literary.

Only the future will tell which is which. Maybe both!

Update the Panton Principles please

There is a big contradiction between the text of The Panton Principles and the List of the Recommended Conformant Licenses. It appears that it is intentional, I’ll explain in a moment why I write this.

This contradiction is very bad for the Open Science movement. That is why, please, update your principles.

Here is the evidence.

1. The second of the Panton Principles is:

“2. Many widely recognized licenses are not intended for, and are not appropriate for, data or collections of data. A variety of waivers and licenses that are designed for and appropriate for the treatment of data are described [here](http://opendefinition.org/licenses#Data). Creative Commons licenses (apart from CCZero), GFDL, GPL, BSD, etc are NOT appropriate for data and their use is STRONGLY discouraged.

*Use a recognized waiver or license that is appropriate for data.* ”

As you can see, the authors clearly state that “Creative Commons licenses (apart from CCZero) … are NOT appropriate for data and their use is STRONGLY discouraged.”

2. However, if you look at the List of Recommended Licenses, surprise:

Creative Commons Attribution Share-Alike 4.0 (CC-BY-SA-4.0) is recommended.

3. The CC-BY-SA-4.0 is important because it has a very clear anti-DRM part:

“You may not offer or impose any additional or different terms or conditions on, or apply any Effective Technological Measures to, the Licensed Material if doing so restricts exercise of the Licensed Rights by any recipient of the Licensed Material.” [source CC 4.0 licence: in Section 2/Scope/a. Licence grant/5]

4. The anti-DRM is not a “must” in the Open Definition 2.1. Indeed, the Open Definition clearly uses “must” in some places and “may” in another places.  See

“2.2.6 Technical Restriction Prohibition

The license may require that distributions of the work remain free of any technical measures that would restrict the exercise of otherwise allowed rights. ”

5. I asked why is this here. Rufus Pollock, one of the authors of The Panton Principles and of the Open Definition 2.1, answered:

“Hi that’s quite simple: that’s about allowing licenses which have anti-DRM clauses. This is one of the few restrictions that an open license can have.”

My reply:

“Thanks Rufus Pollock but to me this looks like allowing as well any DRM clauses. Why don’t include a statement as clear as the one I quoted?”

Rufus:

“Marius: erm how do you read it that way? “The license may prohibit distribution of the work in a manner where technical measures impose restrictions on the exercise of otherwise allowed rights.”

That’s pretty clear: it allows licenses to prohibit DRM stuff – not to allow it. “[Open] Licenses may prohibit …. technical measures …”

Then:

“Marius: so are you saying your unhappy because the Definition fails to require that all “open licenses” explicitly prohibit DRM? That would seem a bit of a strong thing to require – its one thing to allow people to do that but its another to require it in every license. Remember the Definition is not a license but a set of principles (a standard if you like) that open works (data, content etc) and open licenses for data and content must conform to.”

I gather from this exchange that indeed the anti-DRM is not one of the main concerns!

6. So, until now, what do we have? Principles and definitions which aim to regulate what Open Data means which avoid to take an anti-DRM stance. In the same time they strongly discourage the use of an anti-DRM license like CC-BY-4.0. However, on a page which is not as visible they recommend, among others, CC-BY-4.0.

There is one thing to say: “you may use anti-DRM licenses for Open Data”. It means almost nothing, it’s up to you, not important for them. They write that all CC licenses excepting CCZero are bad! Notice that CC0 does not have anything anti-DRM.

Conclusion. This ambiguity has to be settled by the authors. Or not, is up to them. For me this is a strong signal that we witness one more attempt to tweak a well intended  movement for cloudy purposes.

The Open Definition 2.1. ends with:

Richard Stallman was the first to push the ideals of software freedom which we continue.

Don’t say, really? Maybe is the moment for a less ambiguous Free Science.

The price of publishing with GitHub, Figshare, G+, etc

Three years ago I posted The price of publishing with arXiv. If you look at my arXiv articles then you’ll notice that I barely posted on arXiv.org since then. Instead I went into territory which is even less recognized as serious by a big part of academia. I used:

The effects of this choice are put in front of my homepage, so go there to read them. (Besides, it is a good exercise to remember how to click on links and use them, that lost art from the age when internet was free.)

In this post I want to explain what is the price I paid for these choices and what I think now about them.

First, it is a very stressful way of living. I am not joking, as you know stress comes from realizing that there are many choices and one has to choose. Random reward from the social media is addictive. The discovery that there is a way to get out from the situation which keeps us locked into the legacy publishing system (validation). The realization that the problem is not technical but social. A much more cynical view of the undercurrents of the social life of researchers.

The feeling that I can really change the world with my research. The worries that some possible changes might be very dangerous.

The debt I owe concerning the scarcity of my explanations. The effort to show only the aspects I think are relevant, putting aside those who are not. (Btw, if you look at my About page then you’ll read “This blog contains ideas from the future”. It is true because I already pruned the 99% of the paths leading nowhere interesting.)

The desire to go much deeper, the desire to explain once again what and why, to people who seem either lacking long term attention capability or having shallow pet theories.

Is like fishing for Moby Dick.

How to use the chemlambda collection of simulations

The chemlambda_casting folder (1GB) of simulations is now available on Figshare [1].

How to use the chemlambda collection of simulations? Here’s an example. The synthesis from a tape video [2] is reproduced here with a cheap animated gif. The movie records the simulation file 3_tape_long_5346.html which is available for download at [1].

That simple.

If you want to run it in your computer then all you have to do is to download 3_tape_long_5346.html from [1], download from the same place d3.min.js and jquery.min.js (which are there for your convenience). Put the js libs in the same folder as the html file. Open the html file with a browser, strongly recommend Safari or Chrome (not Firefox which blocks with these d3.js animations, for reasons related to d3). In case your computer has problems with the simulation (I used a macbook pro with safari) then slow it like this: edit the html file (with any editor) and look for the line starting with

return 3000 + (4*(step+(Math.random()*

and replace the “4” by “150”, it should be enough.

Here is a longer explanation. The best would be to read carefully the README [4].
“Advanced”: If you want to make another simulation for the same molecule then follow the steps.

1. The molecule used is 3_tape_long_5346.mol which is available at the library of chemlambda molecules [3].

2. So download the content of the gh-pages branch of the chemlambda repository at github [4] as explained in that link.

3. then follow the steps explained there and you’ll get a shiny new 3_tape_long_5346.html which of course may be different in details than the initial one (it depends on the script used, if you use the random rewrites scripts then of course the order of rewrites may be different).

[1] The Chemlambda collection of simulations
https://doi.org/10.6084/m9.figshare.4747390.v1

[2] Synthesis from a tape
https://plus.google.com/+MariusBuliga/posts/Kv5EUz4Mdyp

[3] The library of chemlambda molecules
https://github.com/chorasimilarity/chemlambda-gui/tree/gh-pages/dynamic/mol

[4] Chemlambda repository (readme) https://github.com/chorasimilarity/chemlambda-gui/blob/gh-pages/dynamic/README.md