Arxiv.org is a great resource where many, if not the majority, of results in my field appear long before formal publication in journals. In fact, I don’t have a habit of reading physics journals anymore, but I daily look through new postings on arxiv.I also have two papers submitted only to arxiv, and not to any old-fashioned journal.
That said, arxiv is stuck in the 1990’ies with its focus on lists, TeX, its awkward search and lack of any social network functions. Since it is such a convergence point for the physics community, its backwardness has grown into a serious limiting factor for the free and open scientific process. In other words, what it offers – namely instant publication – is better than what journals offer, but this has in the meantime become the new normal. While what it does not offer is holding us back.
This can be seen in the new massive survey that arxiv has conducted of its own users. In a long list of boring questions about tiny incremental improvements to the website, there is a very important category they called ‘New Services’. You will see that over 55% of survey respondents say that ranking and comment functions, familiar from social networks, reddits, and just, ahem, the entire internet, are either ‘Very Important’ or ‘Somewhat Important’. A smaller majority has just taken the UK out of the EU!
Yet the arxiv program director at Cornell Oya Rieger writes about it as an even split between those who are strongly for these features (~35%) and those strongly against (~35%). She goes on an on about caution and caveats, which basically means that her and the arxiv team are not going to do this on their own. She does mention that the support for these features is stronger among younger users, so there may be a generational divide at play here, and the arxiv team is on the wrong side of this divide from the historical point of view.
Think about it: all of arxiv content is open to the entire internet. If somebody makes a different website which implements these annotation, ranking, search, communication features nicely, and if the community starts using that service, then not having these features as part of arxiv.org itself will be akin to hiding one’s head in the sand – ignoring the new norm that just grew around you. Now, this has not happened yet, but the demand for it is clearly present, as the survey results demonstrate. When this finally happens, it will be the beginning of the end of arxiv, as at that point it will be easier to submit your work to the new website where it can be instantly evaluated, discussed, ranked, categorized and improved through community interactions.