Add a diagonalization section to the references page, and name the executables

[Ooolab]

Adds the missing diagonalization entry to /documentation/pubs/refs/ and tidies the section headings.

New: Exact and full diagonalization — sparsediag and fulldiag

sparsediag and fulldiag ship in every release and are used throughout the ED tutorials, but had no entry on this page. Added directly after the scheduler, covering applications/diag/:

Implementation papers

B. Bauer, L. D. Carr, H. G. Evertz, A. Feiguin, … and S. Wessel, The ALPS project release 2.0: open source software for strongly correlated systems, J. Stat. Mech. 2011, P05001 (2011).

Algorithm papers

C. Lanczos, An iteration method for the solution of the eigenvalue problem of linear differential and integral operators, J. Res. Natl. Bur. Stand. 45, 255-282 (1950).

The ALPS 2.0 entry reuses the existing bauer2011.bib — the homepage Cite button already points at it — rather than duplicating the record. This is the paper's first appearance on the references page itself. lanczos1950.bib is new; its DOI (10.6028/jres.045.026) was checked against the publisher record before being written.

Button counts follow the existing convention: ALPS 2.0 gets Journal / arXiv / Scholar / BibTeX (it has preprint 1101.2646), Lanczos gets Journal / Scholar / BibTeX (1950, predates arXiv) — same as the Swendsen–Wang and Wolff entries.

Dropped the redundant "ALPS" from two headings

Before	After
ALPS Parallel Monte Carlo Scheduler	Parallel Monte Carlo Scheduler
ALPS-DMFT: CT-QMC Impurity Solvers and Dynamical Mean-Field Theory	DMFT: CT-QMC Impurity Solvers and Dynamical Mean-Field Theory

This is the ALPS site's own references page, so the prefix carried no information.

Named the executables in the remaining sections

Every section that has a binary now follows the X — app pattern the rest of the page already used:

Section	Executables
Parallel Monte Carlo Scheduler	(library, none)
Exact and full diagonalization	sparsediag, fulldiag
Classical Monte Carlo for Spin Models	spinmc
Loop Algorithm QMC	looper
The Directed Loop code in SSE representation	dirloop_sse
Worm code	worms
Quantum Wang-Landau Flat-Histogram QMC	qwl
DMFT: CT-QMC Impurity Solvers…	dmft, interaction, hybridization
DMRG	dmrg

Two deliberate omissions:

• hirschfye is left out of the DMFT heading. It builds from the same applications/dmft/qmc/ directory, but Hirsch-Fye is a discrete-time auxiliary-field method — listing it under a heading that says CT-QMC would be inaccurate. The Source: line still covers it.
• Post-processing helpers are omitted (dirloop_sse_evaluate, fulldiag_evaluate), matching how spinmc, qwl and looper already read even though their *_evaluate tools also ship.

Executable names were taken from the ALPS CMakeLists.txt files rather than guessed: sse4 builds dirloop_sse (the current solver), sse holds the legacy dirloop_sse_v1, sse2 builds no executable, and the dmft_* targets under applications/dmft/qmc/ are test binaries.

Applied identically to en, zh-cn and ja — the page is in English on all three, so the files stay byte-identical.

Test plan

• hugo --gc builds clean
• All 23 /data/*.bib targets referenced by the page resolve in the built output
• Rendered section verified: citation lines, author lists and button rows all present
• Lanczos DOI confirmed against the publisher record before writing the BibTeX
• Confirmed no internal anchor links pointed at the two renamed headings, so nothing breaks
• Visual check of the page in a browser

Unrelated pre-existing issue noticed while checking: daley2004.bib, dolfi2014.bib, schollwock2011.bib, vidal2004.bib and white2004.bib sit unreferenced in static/data/, orphaned by the TEBD removal in #100. Left alone here as out of scope.

🤖 Generated with Claude Code

[Ooolab]

Some section titles were changed so that they look consistent. References were from the overleaf table, as @vws100 suggested. Do they look ok?

[egull]

Looks good. I think the Lanczos implementation paper is also a good twist, though the fulldiag code really uses a simple dsyev call, rather than the iterative Lanczos. Let's leave it.