As has been noted in a few places, Neil Withers, one of the editors of soon to be newest Nature journal, Nature Chemistry put out a request last week for input on a range of issues to do with how people use journals, formats, and technical widgets. Egon Willighagen, Rich Apodaca, and Oscar the Journal Munching Robot (masquerading as Peter Murray-Rust, or is that the other way around?) have already posted responses. Here I want to add my own thoughts and possibly amplify some of the points others have made.

From the Nature ‘Sceptical Chymist‘ Blog:

(1) HTML vs PDF: does anyone read the HTML articles? Do you read the PDF on-screen or print it out?

I know I am in the minority here but I read almost exclusively the HTML. Rich says he directly prints the PDF to avoid reading on screen. Egon says he scans the HTML version and then may print a PDF. If I want to print it out, I will use the PDF version, but I almost never print papers out these days. If I need to check something I will go back to the web version. Walking around the write up areas of students at Southampton I think this is also the way they read the papers (yes a lot of them print off a pdf copy, but I’m not sure how many of them actually read it).

Format is dealt with later probably but a good HTML version has to be very clean and clear and make sure it uses the full available window width efficiently. Rich’s point about readability and ease of finding the paper are key. Make it easier and we can save paper. And paper is getting expensive.

(2) Big vs little graphics: what does everyone else think about the tiny size of the graphics in ACS html articles?

It sucks. But its a tough balance, depends on screen real estate etc. If you want to do something really interesting give people the option of using something like the Google Chart API and expose the raw data on the page. Tools exist for making visually appealling interpretive graphics. If you can provide the tools, e.g. as part of a document template that makes it easier for authors to do things this way then you are leading the way to a powerful exposure of raw data.

(3) Tagging/’semantic web’: what do you think about the toys on the RSC’s Project Prospect? What kind of things would you like to see tagged/linked to other content in Nature Chemistry? For instance, Steve would love to do something with named reactions.

I don’t know a lot about Prospect from a technical or user perspective so I can’t comment on implementations etc. But overall there is a real opportunity to do something qualitatively different to the competition. Adopting CML would possibly be a risk but chemistry is a structured science (as well as a structural science) and if this can be made easy for authors (or even invisible to authors) then there is the opportunity for real advances. Semantics is a real chicken and egg issue but if someone takes a lead then the applications will follow in my opinion.

Egon also mentions social tagging and aggregation. Yes please! Obvious things like one click addition to citeulike/connotea but what about access to the names and reference lists of those who have collected the paper (where they are public obviously)? Aggregate blogs and other media that refer back to the paper and link out. Thinking forward – who has used any of these compounds in another paper/protocol/patent. Provide an index of other relevant material. But remember the answer to question 1 and don’t clutter up the page if at all possible. You guys presumably have good interface/graphic design people on staff who can sort out those type of tensions?

(4) 3D molecular structures: do these help your understanding of a paper?

If there is a 3D issue then yes. But how about we start with 2D structures exposed in a sensible way. For every structure a nice exposed 2D picture, with other machine readable names exposed with it. Then a button on the image to play with the 3D structure if desired. Think of this as a test of your system rather than a feature. If the link between compounds, identifiers, and structures is done right, then providing this sort of functionality ought to be relatively easy. Big issue again here is how do you make it easy enough for authors to put all the right information in?

(5) How useful to you are InChIs and SMILES?

At the moment not very, but expose them (not necessarily visibly) so they can be crawled. And CML, and anything else. keep it hidden until someone wants it. Like that 3D button above, provide those when people want them. And anything else you can think of. Converting one to another is pretty trivial where its possible. These will become more and more important as more aggregation and integration services like Chemical Blogspace are put together. And as the tools to put together a personal aggregator and filter come of age.

(6) Forward linking: the RSC and Elsevier/Science Direct offer this – do you use it? Would you use an RSS feed that alerted you to new citations of a particular paper.

I find this often quite useful. But don’t restrict it to papers. Link to all any relevant database entries, blogs, news articles, whatever. As Rich says, providing this in a range of forms (click through to list, RSS, email alerts) will increase uptake. Many chemists are not up on these technologies but use of them will increase and people desperately need new ways to navigate the literature.

(7) Would you actually comment on papers if there was a comments box at the end?

Maybe sometimes, but probably more likely to comment on my blog so as above, trackbacks are valuable. Some will be more comfortable with comments on the site, many will not want to comment at all, and much comment may be elsewhere. As above, aggregate it all and present it in multiple different ways.

(8) We really like the Biochemical Society’s HTML article style (sample one here) – do you?

I can’t comment on the technicalities but as above. Simple. Clear. Clean. Large text. And avoid frames.

Overall, I think there are two themes to my comments. One is that of serving up the same underlying data to different users in different ways according to their likes and dislikes. PDF or HTML, provide both. 2D or 3D, provide both. INCHI/CML/SMILES, yes, yes, and yes, depending on who is asking. The advantage you have is that for much (not all) of chemistry the data can be fitted onto reasonably stable underlying patterns.  So its not as hard as it might be for some fields. I’m not saying its easy but if you want to break the mould it will take some work.

The other underlying theme is how do actually persuade authors to put in the extra information that is required. If you want to provide 2D structure, 3D structure, SMILES, CML, CMLReact etc etc. where do you get this from? The author is not getting any immediate benefit out of putting all this extra data in (except obviously being published in Nature Chemistry) so how do you persuade them to put the effort in? What I talk about above involves re-imagining what the paper is. Yes there will be a printable version, but it will be a rather pale version of the interactive thing to be found on the web. That’s exciting and some people will buy into that but chemistry is inherently conservative so its not enough. The submission process has to be structured so as to make it easy, but that I think is the subject for another post. Besides, Egon said I don’t do enough chemistry to be aggregated on Chemical Blogspace so I need to spread out what I do have :)

Overall I would say the biggest change you could make is to move from an assumption that the PDF is the ‘proper’ version of the paper and to re-think it as a rather pale imitation of the real thing for a somewhat outmoded, but still useful, medium. If the online version provides a readable and useful interactive experience it will be a fundamental shift. And I think one that would be a move forward. You will have to wait for others to catch up, but that’s the price of taking the lead.