Skip to the Main Content

Note:These pages make extensive use of the latest XHTML and CSS Standards. They ought to look great in any standards-compliant modern browser. Unfortunately, they will probably look horrible in older browsers, like Netscape 4.x and IE 4.x. Moreover, many posts use MathML, which is, currently only supported in Mozilla. My best suggestion (and you will thank me when surfing an ever-increasing number of sites on the web which have been crafted to use the new standards) is to upgrade to the latest version of your browser. If that's not possible, consider moving to the Standards-compliant and open-source Mozilla browser.

September 8, 2005

arXiv Physics Reorganization

After long discussions, the Physics Advisory Committee for the arXivs has recommended a reorganization of the physics archives. It’s a fairly drastic overhaul, and you can read about the proposed new categories.

Among the biggest changes:

  • astro-ph has been broken into four categories, based roughly on distance-scale
    • physics.AC: Cosmology
    • physics.AG: Galaxies
    • physics.AS: Stars
    • physics.EP: Earth & Planetary
  • The cond-mat categories have been thoroughly reorganized.
  • hep-ph has been split into
    • physics.HP: General Phenomenology
    • physics.HQ: QCD
  • “Supercategories” have been introduced.
    • Subscribing to physics.SS gets you all of condensed matter and AMO.
    • Subscribing to physics.AA gets you all the astrophysics categories.
    • Subscribing to physics.HH gets you all four high energy physics categories.
    • Subscribing to physics.NN gets you all of nuclear physics.

The form of paper IDs will change to match the new categories. And, yeah, there are a lot more moderators than there were previously.

Your comments on these changes are being solicited, so please look over the changes and offer whatever feedback you feel is needed.

Posted by distler at September 8, 2005 2:14 AM

TrackBack URL for this Entry:

16 Comments & 0 Trackbacks

Re: arXiv Physics Reorganization

does this mean the category names like hep-th and gr-qc are a thing of the past?

Doesn’t this affect searching etc, like I am looking for a paper but I am not sure if it came out before or after the renaming. So I have to search twice, once hep-th and once physics.HT? Or is this going to be implemented retroactively so all papers of the past are reassigned to the new categories? What happens to everybody’s homepages, cv’s etc, do they have to be updated?

In principle, adjusting categories to the need sounds a like a good idea but what about all these transition problems?

being in conservative mood today.

Posted by: Robert on September 8, 2005 3:33 AM | Permalink | Reply to this

Re: arXiv Physics Reorganization

The Math archives underwent a similar reorganization years ago. alg-geom became math.AG. This does not, in my experience, diminish your ability to search for old papers.

But don’t take my word for it, go to the search page and try it out for yourself.

Posted by: Jacques Distler on September 8, 2005 7:01 AM | Permalink | PGP Sig | Reply to this

Re: arXiv Physics Reorganization

I think it’s a great idea to reorganize some of the categories, in particular those that have been less accepted than hep-th. I remember vividly how a condensed matter colleague of mine said, “I used to post all my papers on cond-mat, but I’ve kind of stopped doing that”. I don’t know what the reason is – maybe string theorists are just smarter than other physicists? ;) –, but the same would *never* happen to hep-th.

That said, I’m kind of sad that they also want to change hep-th, because there is absolutely no reason for that! Never change a running system…


Posted by: Michael on September 8, 2005 9:43 AM | Permalink | Reply to this

Re: arXiv Physics Reorganization

The naming strikes me as a bit odd. You seem to have settled for 10-letter names, but then you use almost all of the available code space to denote math vs. physics (rather general information). All of the interesting information is crammed into the last two letters. And, lets face it, the two lettter codes are not self-explanatory. If space were at a premium, why not use pGR insted of physics.GR?

What is wrong with a usenet-like hierachical and human readable naming scheme, like


If chosen carefully, one could then uniquely abbreviate them as p.a.c, p.a.g, p.a.s, p.a.p.

The supercategories are good, but even better would be a dynamically generated homepage with the archives of your choice.

Posted by: Volker Braun on September 8, 2005 10:02 AM | Permalink | Reply to this

Naming Conventions

If chosen carefully, one could then uniquely abbreviate them as p.a.c, p.a.g, p.a.s, p.a.p.

And, in fact, we’ve done precisely that, except that we re-expanded the initial “physics”: physics.AC, physics.AG, physics.AS. (Earth and Planetary Physics encompasses geophysics and some planetary astronomy, following a trend in the field; it’s not really a “sub-category” of astro.)

I, personally, happened to like the longer abbreviations (hep-th versus physics.HT), but I can see the virtues of a more uniform naming convention.

Note that, unlike USENET, there is no hierarchy here. Despite the existence of supercategories, you can’t submit a paper to one of these higher-level aggregations.

The supercategories are good, but even better would be a dynamically generated homepage with the archives of your choice.

That, too, is coming. If you want to subscribe to physics.AC, physics.HP and physics.HT, you’ll be able to create your own “supercategory”. The pre-defined supercategories were created for those unwilling or unable to customize their arXiv experience for themselves.

Posted by: Jacques Distler on September 8, 2005 10:32 AM | Permalink | PGP Sig | Reply to this

Re: arXiv Physics Reorganization

Actually the main function of the supercategories is not to cater to the unwilling or the unable. The arXiv has never had much loyalty to that group; although things will hopefully get much better, it will remain a secondary concern.

The real point of the supercategories is to reassure authors that readers will use those supercategories. It would be disruptive, and otherwise a source of FUD, if authors had to guess all of the supercategories that readers cobbled together. As Jacques siad, readers will still be able to do that, but it is expected to be a side issue. Predefined supercategories are part of the strong coupling between authors and readers that sustains the arXiv.

This model is supported by experimental data, in that users expressed fear that cond-mat and astro-ph might “go away”.

Posted by: Greg Kuperberg on September 8, 2005 1:54 PM | Permalink | Reply to this

Re: arXiv Physics Reorganization

It’s true that many people have expressed that concern. The part that I have had a hard time sorting out is the extent to which they are expressing a concern as readers (I will no longer be able to easily see all abstracts in Condensed Matter Physics) or as authors (stellar astronomers will no longer automatically see my latest paper about polarization in the CMBR).

From the author’s point of view, it’s somehow reassuring that someone, who would not ordinarily subscribe to physics.AC, would nonetheless see your paper because they would subscribe to the physics.AA supercategory.

Posted by: Jacques Distler on September 8, 2005 2:24 PM | Permalink | PGP Sig | Reply to this

Re: arXiv Physics Reorganization

“as authors”

That is, every user who has said this is both an author and a reader. Regardless of what they meant, and accepting your semantics, the logical interpretation is mainly “as authors” and somewhat “as readers”. physics.AA helps assure authors of the viability of components like physics.AC. Moreover, if many authors assume a physics.AA audience, then some readers will want to read physics.AA to detect all papers by those authors.

The dynamical coupling between authors and readers is a fundamental property of the arXiv. It is part of the answer to many questions, like why the arXiv works better than home pages + CiteSeer, why arXiv moderation isn’t hopeless, etc.

Posted by: Greg Kuperberg on September 8, 2005 5:47 PM | Permalink | Reply to this

Re: arXiv Physics Reorganization

The category choices look sensible to me, but I’m nonplussed by the naming convention, for two reasons (partially echoing Volker’s and Jacques’s comments above):

1) The new two-character names are harder to understand than the old ones – is there some technical constraint preventing the use of longer names?

2) If I am looking at a category name there is no semantic clue that tells me whether it is a supercategory or not. This feels wrong, since the two types of category obey different rules, e.g. one can post papers to physics.AC but not to physics.AA.

Posted by: Andrew Neitzke on September 9, 2005 12:51 AM | Permalink | Reply to this

Re: arXiv Physics Reorganization

Subscribing to a supercategory to read a range of preprints seems like a good idea, but what is the rationale for the groupings? Isn’t there more overlap between say gr-qc and hep-th than hep-th and gr-qc certainly cross-lists to hep-th quite a bit at the moment.

Posted by: boreds on September 9, 2005 4:22 AM | Permalink | Reply to this


…what is the rationale for the groupings? Isn’t there more overlap between say gr-qc and hep-th than hep-th and gr-qc certainly cross-lists to hep-th quite a bit at the moment.

The physics.AA and physics.SS supercategories, obviously, represent purported existing communities of readers (astro-ph and cond-mat).

In the case of nuclear physics, there was a certain amount of debate as to whether to merge nucl-th and nucl-ex. In the end, the consensus was to keep them separate, but to create a supercategory encompassing them both.

The high energy physics supercategory, physics.HH was certainly the odd duck. The hep-th readers were assumed to also be interested in hep-ph. hep-ph readers were assumed to also be readers of hep-ex. And hep-ex readers were assume to be interested in physics.ID (as were the observational astronomers and the nuclear and condensed-matter experimentalists).

In reality, it is far more likely that a hep-th reader is interested in gr-qc than in physics.ID.

[In fact, not much thought was put into the reading habits of existing hep-\* readers. physics.HH is simply the union of all physics.H\* categories plus physics.ID.]

So I’m gonna assume that, when the option becomes available to them, most readers are going to subscribe to supercategories of their own design, rather than to the pre-defined ones.

I’m not convinced that the implications of this for authors are really more dire than if readers stuck to existing supercategories.

By the way, when I say “subscribe”, I mean either to a daily emailings, to an rss feed, or simply visiting a customized web page, listing new abstracts in the chosen (super)category.

Posted by: Jacques Distler on September 10, 2005 2:34 AM | Permalink | PGP Sig | Reply to this

Re: arXiv Physics Reorganization

Frankly I don’t think that it is a terribly good idea. It will create chaos, maybe even invalidate web links that point to the articles (which would be really bad), and so forth. Do you really want to change these things retroactively?

If yes, how will you sort all the articles? If not, will the two conventions co-exist? Strange in any case.

A part of my negative attitude may be just because the new names look so ugly (and even those for which the dot-abbreviation format has been used still look ugly to me). Why don’t you follow the standard arXiv hep-th-like conventions?

That would mean something like:

astro-cosm, astro-galax, astro-star, astro-planet, hep-model, hep-qcd, phys-cm, phys-as, phys-he, phys-np

You may also subdivide hep-th to hep-st (string theory) and hep-ns (no strings).

Posted by: Lubos Motl on September 11, 2005 2:05 PM | Permalink | Reply to this

Re: arXiv Physics Reorganization

Also, when you already changing the formats, I would propose to introduce a unified number identification of every new preprint from any arxiv, such as 0510140. Because there are more than 1000 papers monthly, they would continue like 0509A00 after 0509999. Is not 3600 papers a month enough? If not, there are obvious ways how to extend the sequence using two letters, such as 0509AA0, and so forth. It could simplify life if preprints could be parameterized by seven characters only. Of course, the hep-qcd things could be added, but not as a necessary condition.

Posted by: Lubos Motl on September 11, 2005 2:12 PM | Permalink | Reply to this

Re: arXiv Physics Reorganization

I would strongly urge that the existing names be kept and new ones added in the same style. hep-th, hep-ph. gr-qc are all embedded in our mental habits and they feel good. Why change them? Yes, they could be more uniform, but the new names physic.ab are very untransparent and feel more like something librarians would invent than physicists. The original names arose at a moment when the archive was new and less formal and are hence more reflective of the style and culture of physicists. There is also value in keeping names for things which we have become familar and happy with.

On a more substantial note, there has been for some time an ambituity as to whether non-string quantum gravity papers go to hep-th or gr-qc. Most that are not explicitly quantum cosmology go to hep-th and this is a good thing as it means papers on all the diffferent approaches to quantum gravity are in the same place. This matters because the alternative is that people who work on different approaches to the same problem become ignorant about what the others are doing.

I would thus propose to formalize this and move quantum gravity generally to
physics.HT, labeling it
High-Energy – Field Theory, String Theory and Quantum Gravity



Posted by: Lee Smolin on September 14, 2005 7:26 PM | Permalink | Reply to this

Straddling the boundaries

It is inevitable that no matter where you draw the boundaries between categories, some topics will straddle the boundaries between categories.

There will inevitably be papers on the quark-gluon plasma in nucl-th (now physics.NT) and in hep-ph (now physics.HQ). There will be papers on brane-worlds in hep-th (physics.HT) and in hep-ph (physics.HP), and so on. This is unavoidable.

Anyone seriously interested in a topic which straddles the boundary between two categories (in the case at hand, hep-th and gr-qc) will surely subscribe to both and will, in all likelihood, cross-list their papers.

gr-qc was created in July 1992 at the request of many of the same people whose work you would now move back to hep-th (physics.HT). Whatever the intellectual merits, there was a perceived sociological need to create a separate archive. I think that, to a great extent, we (on the committee) need to bow to the sociological realities about existing communities of scholars, rather than try to impose a structure which we, personally, think would have greater intellectual merit.

There is also the issue of size. astro-ph was unmanageably huge, and was split in 3. hep-ph was also thought to be too large (and, with the imminence of the LHC, destined to grow yet larger), and was split in 2. hep-th was felt to be still (barely) of a manageable size. However, merging all or most of gr-qc back into hep-th would produce an arXiv too unwieldy to moderate effectively … in other words, one that would be ripe to be split up.

Posted by: Jacques Distler on September 16, 2005 10:41 AM | Permalink | PGP Sig | Reply to this

Re: arXiv Physics Reorganization

Didn’t the ArXiv people try to enforce a similar change around 1997? Everything should go to the physics archive, with different subject classes; hep-th and perhaps gr-qc were kept for backward compatibility only. Evidently people didn’t like this order, because the policy changed in 1998. In particular, all posts to physics, Subj. class Mathematical physics were automatically cross-posted to math-ph.

Posted by: Thomas Larsson on September 15, 2005 5:47 AM | Permalink | Reply to this

Post a New Comment