What are the best practices on publishing preprints and e-prints of your results?
The standard at my current workgroup is the following:
- make a preprint at our preprint server as soon as it's done, "to put a timestamp on it"
- make an e-print at our e-print server just before publishing, "to have all the work of the group available in the same place, and accessible in a reasonably final version without paying license fees"
In addition, I believe in the utility of arXiv as a "global" e-print server, and I try to put my papers there, too. However, this whole process seems redundant to me. I think that arXiv could replace 1 and 2 without many drawbacks. I am a bit perplexed about the necessity to use two separate servers (and two sets of metadata to update manually).
How is the "publishing history" of your results? Which preprint or e-print servers do you use? What are the important points to care about when making preliminary versions available?
Additionally, it would be interesting to know how much of the "tedious work" of data entering can actually be automated:
In how many different places do you update the metadata for your papers manually (including your website and CV)? Is there a smart way have a preprint/eprint server do your work and avoid typing in title, authors and journal number about ten different times?