Egothor/Radixor
2
0
Fork 0
Code Issues Pull Requests Packages Projects Releases Wiki Activity
Files
9c5b9e331bd1cf2fe67b8062206a0685f868e2d9
Radixor/docs/benchmarks/languages/russian.md
Leo Galambos 05f3855b99 feat(benchmarks): expand multilingual stemming quality evaluation 
* cover all Radixor dictionary languages
* add PRIMARY_OUTPUT, ANY_CANDIDATE, and ALL_CANDIDATES policies
* measure pairwise over-stemming and under-stemming
* add balanced accuracy and complementary quality metrics
* compare single-output and multi-output stemmers fairly
* improve result validation, reporting, and documentation
* move stemming quality tests into the standard test source set
* preserve the existing JMH benchmark structure and badge output
2026-07-20 23:20:17 +02:00

26 KiB
Raw Blame History

Russian Stemmer Benchmarks

This page reports same-language stemming benchmarks for Russian. Accuracy is listed first because speed without root agreement is not enough to interpret stemmer quality.

All speed values are environment-specific and were measured on the hardware and JVM listed in the benchmark overview. Speed benchmark operations process changed dictionary tokens only. Accuracy uses the complete Radixor dictionary for the language.

Radixor must not be read as simply "slower" when a narrow competitor has a lower timing row. In these tables Radixor is the quality-oriented baseline: its exact-root accuracy is typically close to 100%, while many faster rule-based, light, minimal, or possessive filters reach that speed by doing much less linguistic work and often score far lower in All exact and Changed exact. The Radixor rows in this benchmark refresh use the contracted compiled patch trie: compilation collapses uniform patch-command subtrees into accepting leaves, reducing hot lookup depth while preserving the preferred stemming result measured by the accuracy pass. The EnglishRadixorDictionaryCoverageBenchmark table shows the resulting quality/speed envelope explicitly. The same interpretation applies to this language page: speed rows must be read together with the accuracy table above them.

Dictionary Corpus

Resource Dictionary rows Complete quality tokens Already-root tokens Changed speed tokens
RU_RU 37,410 806,279 74,808 731,471

Radixor Patch Command Distribution

Radixor stores the preferred transformation for each normalized dictionary word form as a compiled patch command. This distribution shows which runtime command class is selected by the trained trie for the complete language dictionary. The total number of preferred patch commands analyzed for this language is 806,279.

Command class Meaning Word forms Share
AppendCharacterCommand Appends one character to the end of the word form. 9,260 1.148%
BackwardCompoundCommand Applies a multi-step backward patch made from skip, delete, insert, and replace operations. 584,785 72.529%
DeleteSuffixCommand Deletes one or more trailing characters from the word form. 82,864 10.277%
PreserveCommand Returns the word form unchanged because it already matches the preferred root. 75,646 9.382%
ReplaceLastCharacterCommand Replaces the final character of the word form. 53,724 6.663%

Accuracy

Accuracy is computed from JMH auxiliary counters in the current report. The counters are deterministic for a fixed corpus and stemmer; percentages divide matching counters by evaluated counters from the same report and are not timing metrics.

Stemmer All exact Changed exact Root preserved Note
Radixor 98.807% 98.696% 99.896% Full Radixor dictionary patch-command stemmer.
Lucene RussianLightStemFilter 9.658% 8.452% 21.447% Light suffix stemmer; intentionally narrower than a dictionary-derived stemmer.
Lucene SnowballFilter 9.162% 8.162% 18.936% Lucene TokenFilter integration path around the Snowball algorithm.
Official Snowball direct 9.162% 8.162% 18.936% Official Snowball generated Java stemmer; rule-based suffix algorithm.

Speed

Speed uses JMH average time, 3 warmup iterations, 5 measurement iterations, 1 fork, and 1 thread. Relative factor is computed against the single Radixor row on this language page. Values below 1.000 are faster than that Radixor baseline; values above 1.000 are slower.

Stemmer Benchmark method Score ms/op Error ms ns/token Relative vs Radixor Note
Radixor russianRadixor 89.671 3.886 122.6 1.000 Full Radixor dictionary patch-command stemmer.
Lucene RussianLightStemFilter russianLuceneRussianLightStemFilter 60.522 5.310 82.7 0.675 Light Russian suffix stemmer.
Official Snowball direct snowballDirect[RUSSIAN] 106.031 9.287 145.0 1.182 Official Snowball generated Java stemmer; direct API.
Lucene SnowballFilter luceneSnowballFilter[RUSSIAN] 137.512 10.801 188.0 1.534 Lucene TokenFilter path around Snowball; includes TokenStream overhead.

Interpretation Notes

  • Radixor is a dictionary-derived patch-command stemmer. Its quality depends on the language resource used to train the compiled trie.
  • Light, minimal, plural, and possessive filters are narrow baselines. They can be fast because they intentionally perform less linguistic work.
  • Lucene TokenFilter rows include TokenStream, attribute, and required normalization overhead. Direct rows measure exposed direct APIs.
  • Morfologik rows are dictionary-based and can emit multiple terms for one input token. Quality rows use the first returned term when no ranking weight is available.
  • Snowball rows are rule-based generated suffix stemmers; they are useful algorithmic baselines, not dictionary-root equivalence guarantees.

Stemming Quality

Runtime performance and linguistic grouping quality are independent dimensions. This section evaluates language RU_RU using the complete validated stemming-quality result matrix. Every usable dictionary row is one gold-standard group of forms expected to share a morphological family or lemma. Exact equality with a predetermined lemma is not required. Same-row pairs are positive pairs; pairs from different rows are negative pairs.

ALL_WORDS includes every valid group and its original forms. LOWERCASE_GROUPS_ONLY excludes an entire group when any Unicode code point is uppercase or titlecase; retained words are not lowercased or otherwise rewritten. This isolates case-handling effects without changing retained inputs. Download the complete machine-readable result snapshot.

Evaluation Scope and Key Findings

The dictionary resource is src/main/resources/ru_ru/stemmer.gz. The following findings compare only deterministic PRIMARY_OUTPUT rows over identical included groups; candidate policies are reported separately as capability analyses.

  • ALL_WORDS: Radixor ranks first by balanced accuracy at 0.989827 among 4 deterministic stemmers. The runner-up is SNOWBALL RUSSIAN LUCENE FILTER at 0.834876, a difference of 0.154951. This rank does not imply leadership in throughput or every secondary metric.
  • LOWERCASE_GROUPS_ONLY: Radixor ranks first by balanced accuracy at 0.989852 among 4 deterministic stemmers. The runner-up is SNOWBALL RUSSIAN DIRECT at 0.834854, a difference of 0.154998. This rank does not imply leadership in throughput or every secondary metric.

ALL_WORDS

This mode contains 6 result rows, 4 evaluated stemmers, and 3 output policies. Applied-row and form counts are shown per row because adapters share the language corpus but policy rows remain independently auditable. Rankings are separated by output policy and ordered by unrounded balanced accuracy, followed by MCC, F1, over-stemming rate, over-stemming count, under-stemming rate, and stemmer. Balanced accuracy is a navigation metric, not a universally authoritative quality score.

PRIMARY_OUTPUT ranking

Rank Stemmer Output policy Balanced accuracy Over-stemming Under-stemming F0.5 F1 MCC
1 Radixor PRIMARY_OUTPUT 0.989827 155850 / 295576291016 (0.000053%) 266302 / 13089505 (2.034470%) 0.986313 0.983806 0.983814
2 SNOWBALL RUSSIAN LUCENE FILTER PRIMARY_OUTPUT 0.834876 3785790 / 295576291016 (0.001281%) 4322616 / 13089505 (33.023525%) 0.692485 0.683786 0.683923
3 SNOWBALL RUSSIAN DIRECT PRIMARY_OUTPUT 0.834867 3782908 / 295576291016 (0.001280%) 4322849 / 13089505 (33.025305%) 0.692603 0.683851 0.683989
4 RUSSIAN LUCENE RUSSIAN LIGHT STEM FILTER PRIMARY_OUTPUT 0.617692 321183 / 295576291016 (0.000109%) 10008438 / 13089505 (76.461547%) 0.577011 0.373649 0.461687
Classification metrics
Rank Stemmer Output policy Precision Recall Specificity Balanced accuracy Pairwise accuracy Error rate
1 Radixor PRIMARY_OUTPUT 0.987992 0.979655 0.999999 0.989827 0.999999 0.000001
2 SNOWBALL RUSSIAN LUCENE FILTER PRIMARY_OUTPUT 0.698408 0.669765 0.999987 0.834876 0.999973 0.000027
3 SNOWBALL RUSSIAN DIRECT PRIMARY_OUTPUT 0.698563 0.669747 0.999987 0.834867 0.999973 0.000027
4 RUSSIAN LUCENE RUSSIAN LIGHT STEM FILTER PRIMARY_OUTPUT 0.905597 0.235385 0.999999 0.617692 0.999965 0.000035
Pair-relation metrics
Rank Stemmer Output policy F0.5 F1 F2 Jaccard Fowlkes–Mallows MCC
1 Radixor PRIMARY_OUTPUT 0.986313 0.983806 0.981311 0.968128 0.983815 0.983814
2 SNOWBALL RUSSIAN LUCENE FILTER PRIMARY_OUTPUT 0.692485 0.683786 0.675304 0.519510 0.683936 0.683923
3 SNOWBALL RUSSIAN DIRECT PRIMARY_OUTPUT 0.692603 0.683851 0.675318 0.519585 0.684003 0.683989
4 RUSSIAN LUCENE RUSSIAN LIGHT STEM FILTER PRIMARY_OUTPUT 0.577011 0.373649 0.276278 0.229747 0.461696 0.461687
Partition metrics (PRIMARY_OUTPUT only)
Rank Stemmer Output policy Adjusted Rand Index Homogeneity Completeness V-measure Normalized mutual information
1 Radixor PRIMARY_OUTPUT 0.983805 0.997699 0.997274 0.997487 0.997487
2 SNOWBALL RUSSIAN LUCENE FILTER PRIMARY_OUTPUT 0.683773 0.974131 0.953674 0.963794 0.963794
3 SNOWBALL RUSSIAN DIRECT PRIMARY_OUTPUT 0.683838 0.974180 0.953661 0.963811 0.963811
4 RUSSIAN LUCENE RUSSIAN LIGHT STEM FILTER PRIMARY_OUTPUT 0.373638 0.994311 0.870888 0.928516 0.928516
Raw pair counts
Rank Stemmer Output policy TP FP FN TN Over error / possible Under error / possible
1 Radixor PRIMARY_OUTPUT 12823203 155850 266302 295576135166 155850 / 295576291016 266302 / 13089505
2 SNOWBALL RUSSIAN LUCENE FILTER PRIMARY_OUTPUT 8766889 3785790 4322616 295572505226 3785790 / 295576291016 4322616 / 13089505
3 SNOWBALL RUSSIAN DIRECT PRIMARY_OUTPUT 8766656 3782908 4322849 295572508108 3782908 / 295576291016 4322849 / 13089505
4 RUSSIAN LUCENE RUSSIAN LIGHT STEM FILTER PRIMARY_OUTPUT 3081067 321183 10008438 295575969833 321183 / 295576291016 10008438 / 13089505

ANY_CANDIDATE ranking

Rank Stemmer Output policy Balanced accuracy Over-stemming Under-stemming F0.5 F1 MCC
1 Radixor ANY_CANDIDATE 1.000000 0 / 295576291016 (0.000000%) 13 / 13089505 (0.000099%) 1.000000 1.000000 1.000000
Classification metrics
Rank Stemmer Output policy Precision Recall Specificity Balanced accuracy Pairwise accuracy Error rate
1 Radixor ANY_CANDIDATE 1.000000 0.999999 1.000000 1.000000 1.000000 0.000000
Pair-relation metrics
Rank Stemmer Output policy F0.5 F1 F2 Jaccard Fowlkes–Mallows MCC
1 Radixor ANY_CANDIDATE 1.000000 1.000000 0.999999 0.999999 1.000000 1.000000
Partition metrics (PRIMARY_OUTPUT only)
Rank Stemmer Output policy Adjusted Rand Index Homogeneity Completeness V-measure Normalized mutual information
1 Radixor ANY_CANDIDATE n/a n/a n/a n/a n/a
Raw pair counts
Rank Stemmer Output policy TP FP FN TN Over error / possible Under error / possible
1 Radixor ANY_CANDIDATE 13089492 0 13 295576291016 0 / 295576291016 13 / 13089505

ALL_CANDIDATES ranking

Rank Stemmer Output policy Balanced accuracy Over-stemming Under-stemming F0.5 F1 MCC
1 Radixor ALL_CANDIDATES 0.999999 434710 / 295576291016 (0.000147%) 13 / 13089505 (0.000099%) 0.974119 0.983665 0.983796
Classification metrics
Rank Stemmer Output policy Precision Recall Specificity Balanced accuracy Pairwise accuracy Error rate
1 Radixor ALL_CANDIDATES 0.967857 0.999999 0.999999 0.999999 0.999999 0.000001
Pair-relation metrics
Rank Stemmer Output policy F0.5 F1 F2 Jaccard Fowlkes–Mallows MCC
1 Radixor ALL_CANDIDATES 0.974119 0.983665 0.993401 0.967856 0.983797 0.983796
Partition metrics (PRIMARY_OUTPUT only)
Rank Stemmer Output policy Adjusted Rand Index Homogeneity Completeness V-measure Normalized mutual information
1 Radixor ALL_CANDIDATES n/a n/a n/a n/a n/a
Raw pair counts
Rank Stemmer Output policy TP FP FN TN Over error / possible Under error / possible
1 Radixor ALL_CANDIDATES 13089492 434710 13 295575856306 434710 / 295576291016 13 / 13089505

Multi-output analysis

Alternative candidates are capability analyses, not replacements for the deterministic comparison.

Stemmer Under pairs repaired Best-case over pairs avoided All-candidate collisions added Multi-candidate forms Multi-candidate share Maximum candidates Total candidate assignments
Radixor 266289 155850 278860 19162 2.492190% 4 788492

LOWERCASE_GROUPS_ONLY

This mode contains 6 result rows, 4 evaluated stemmers, and 3 output policies. Applied-row and form counts are shown per row because adapters share the language corpus but policy rows remain independently auditable. Rankings are separated by output policy and ordered by unrounded balanced accuracy, followed by MCC, F1, over-stemming rate, over-stemming count, under-stemming rate, and stemmer. Balanced accuracy is a navigation metric, not a universally authoritative quality score.

PRIMARY_OUTPUT ranking

Rank Stemmer Output policy Balanced accuracy Over-stemming Under-stemming F0.5 F1 MCC
1 Radixor PRIMARY_OUTPUT 0.989852 155850 / 295000681652 (0.000053%) 265613 / 13087126 (2.029575%) 0.986322 0.983830 0.983838
2 SNOWBALL RUSSIAN DIRECT PRIMARY_OUTPUT 0.834854 3782908 / 295000681652 (0.001282%) 4322407 / 13087126 (33.027931%) 0.692561 0.683815 0.683953
3 SNOWBALL RUSSIAN LUCENE FILTER PRIMARY_OUTPUT 0.834854 3782908 / 295000681652 (0.001282%) 4322407 / 13087126 (33.027931%) 0.692561 0.683815 0.683953
4 RUSSIAN LUCENE RUSSIAN LIGHT STEM FILTER PRIMARY_OUTPUT 0.617630 318921 / 295000681652 (0.000108%) 10008238 / 13087126 (76.473918%) 0.577038 0.373540 0.461703
Classification metrics
Rank Stemmer Output policy Precision Recall Specificity Balanced accuracy Pairwise accuracy Error rate
1 Radixor PRIMARY_OUTPUT 0.987991 0.979704 0.999999 0.989852 0.999999 0.000001
2 SNOWBALL RUSSIAN DIRECT PRIMARY_OUTPUT 0.698516 0.669721 0.999987 0.834854 0.999973 0.000027
3 SNOWBALL RUSSIAN LUCENE FILTER PRIMARY_OUTPUT 0.698516 0.669721 0.999987 0.834854 0.999973 0.000027
4 RUSSIAN LUCENE RUSSIAN LIGHT STEM FILTER PRIMARY_OUTPUT 0.906139 0.235261 0.999999 0.617630 0.999965 0.000035
Pair-relation metrics
Rank Stemmer Output policy F0.5 F1 F2 Jaccard Fowlkes–Mallows MCC
1 Radixor PRIMARY_OUTPUT 0.986322 0.983830 0.981350 0.968175 0.983839 0.983838
2 SNOWBALL RUSSIAN DIRECT PRIMARY_OUTPUT 0.692561 0.683815 0.675288 0.519544 0.683967 0.683953
3 SNOWBALL RUSSIAN LUCENE FILTER PRIMARY_OUTPUT 0.692561 0.683815 0.675288 0.519544 0.683967 0.683953
4 RUSSIAN LUCENE RUSSIAN LIGHT STEM FILTER PRIMARY_OUTPUT 0.577038 0.373540 0.276152 0.229664 0.461713 0.461703
Partition metrics (PRIMARY_OUTPUT only)
Rank Stemmer Output policy Adjusted Rand Index Homogeneity Completeness V-measure Normalized mutual information
1 Radixor PRIMARY_OUTPUT 0.983829 0.997697 0.997321 0.997509 0.997509
2 SNOWBALL RUSSIAN DIRECT PRIMARY_OUTPUT 0.683802 0.974149 0.953634 0.963782 0.963782
3 SNOWBALL RUSSIAN LUCENE FILTER PRIMARY_OUTPUT 0.683802 0.974149 0.953634 0.963782 0.963782
4 RUSSIAN LUCENE RUSSIAN LIGHT STEM FILTER PRIMARY_OUTPUT 0.373528 0.994350 0.870767 0.928464 0.928464
Raw pair counts
Rank Stemmer Output policy TP FP FN TN Over error / possible Under error / possible
1 Radixor PRIMARY_OUTPUT 12821513 155850 265613 295000525802 155850 / 295000681652 265613 / 13087126
2 SNOWBALL RUSSIAN DIRECT PRIMARY_OUTPUT 8764719 3782908 4322407 294996898744 3782908 / 295000681652 4322407 / 13087126
3 SNOWBALL RUSSIAN LUCENE FILTER PRIMARY_OUTPUT 8764719 3782908 4322407 294996898744 3782908 / 295000681652 4322407 / 13087126
4 RUSSIAN LUCENE RUSSIAN LIGHT STEM FILTER PRIMARY_OUTPUT 3078888 318921 10008238 295000362731 318921 / 295000681652 10008238 / 13087126

ANY_CANDIDATE ranking

Rank Stemmer Output policy Balanced accuracy Over-stemming Under-stemming F0.5 F1 MCC
1 Radixor ANY_CANDIDATE 1.000000 0 / 295000681652 (0.000000%) 0 / 13087126 (0.000000%) 1.000000 1.000000 1.000000
Classification metrics
Rank Stemmer Output policy Precision Recall Specificity Balanced accuracy Pairwise accuracy Error rate
1 Radixor ANY_CANDIDATE 1.000000 1.000000 1.000000 1.000000 1.000000 0.000000
Pair-relation metrics
Rank Stemmer Output policy F0.5 F1 F2 Jaccard Fowlkes–Mallows MCC
1 Radixor ANY_CANDIDATE 1.000000 1.000000 1.000000 1.000000 1.000000 1.000000
Partition metrics (PRIMARY_OUTPUT only)
Rank Stemmer Output policy Adjusted Rand Index Homogeneity Completeness V-measure Normalized mutual information
1 Radixor ANY_CANDIDATE n/a n/a n/a n/a n/a
Raw pair counts
Rank Stemmer Output policy TP FP FN TN Over error / possible Under error / possible
1 Radixor ANY_CANDIDATE 13087126 0 0 295000681652 0 / 295000681652 0 / 13087126

ALL_CANDIDATES ranking

Rank Stemmer Output policy Balanced accuracy Over-stemming Under-stemming F0.5 F1 MCC
1 Radixor ALL_CANDIDATES 0.999999 434710 / 295000681652 (0.000147%) 0 / 13087126 (0.000000%) 0.974115 0.983663 0.983794
Classification metrics
Rank Stemmer Output policy Precision Recall Specificity Balanced accuracy Pairwise accuracy Error rate
1 Radixor ALL_CANDIDATES 0.967851 1.000000 0.999999 0.999999 0.999999 0.000001
Pair-relation metrics
Rank Stemmer Output policy F0.5 F1 F2 Jaccard Fowlkes–Mallows MCC
1 Radixor ALL_CANDIDATES 0.974115 0.983663 0.993401 0.967851 0.983794 0.983794
Partition metrics (PRIMARY_OUTPUT only)
Rank Stemmer Output policy Adjusted Rand Index Homogeneity Completeness V-measure Normalized mutual information
1 Radixor ALL_CANDIDATES n/a n/a n/a n/a n/a
Raw pair counts
Rank Stemmer Output policy TP FP FN TN Over error / possible Under error / possible
1 Radixor ALL_CANDIDATES 13087126 434710 0 295000246942 434710 / 295000681652 0 / 13087126

Multi-output analysis

Alternative candidates are capability analyses, not replacements for the deterministic comparison.

Stemmer Under pairs repaired Best-case over pairs avoided All-candidate collisions added Multi-candidate forms Multi-candidate share Maximum candidates Total candidate assignments
Radixor 265613 155850 278860 18991 2.472358% 4 787549

Output Policies and Metric Definitions

PRIMARY_OUTPUT uses one deterministic stem per form and therefore defines a strict partition. ANY_CANDIDATE is an optimistic oracle-assisted pairwise upper bound: a same-group pair succeeds when candidates intersect, while a different-group pair succeeds when a non-colliding selection exists. Candidate choices may differ between pairs, so this is not deterministic runtime behaviour and need not represent one globally consistent assignment. ALL_CANDIDATES activates every returned candidate; forms are related when candidate sets intersect. Alternatives can reduce under-stemming but can introduce cross-group collisions, and the resulting relation can overlap and need not be a partition.

For each row, TP = underPossiblePairs - underErrorPairs, FN = underErrorPairs, FP = overErrorPairs, and TN = overPossiblePairs - overErrorPairs. TP and FN concern same-group pairs; FP and TN concern different-group pairs. Consequently, under-stemming and over-stemming use different denominators. Undefined values are rendered as n/a.

  • Under-stemming rate: FN / (TP + FN), the false-negative rate over same-group pairs.
  • Over-stemming rate: FP / (TN + FP), the false-positive rate over different-group pairs.
  • Pairwise precision: TP / (TP + FP), the fraction of predicted conflations that are gold-standard positive pairs.
  • Pairwise recall: TP / (TP + FN), the fraction of gold-standard positive pairs successfully connected.
  • Pairwise specificity: TN / (TN + FP), the fraction of different-group pairs correctly separated.
  • Balanced accuracy: (recall + specificity) / 2. It gives equal weight to positive and negative pair classes and is less dominated by the large true-negative class than ordinary accuracy. It does not replace the raw errors or other metrics.
  • Pairwise F-beta: ((1 + betaSquared) * TP) / (((1 + betaSquared) * TP) + (betaSquared * FN) + FP). F0.5 emphasizes precision and penalizes over-stemming more; F1 weights precision and recall equally; F2 emphasizes recall and penalizes under-stemming more.
  • MCC: (TP * TN - FP * FN) / sqrt((TP + FP) * (TP + FN) * (TN + FP) * (TN + FN)). It uses all confusion counts and remains useful under class imbalance, except when its denominator is degenerate.
  • Jaccard index: TP / (TP + FP + FN).
  • Fowlkes–Mallows index: sqrt(precision * recall).
  • Pairwise accuracy: (TP + TN) / (TP + TN + FP + FN). It can be dominated by true-negative cross-group pairs.
  • Pairwise error rate: (FP + FN) / (TP + TN + FP + FN).

Adjusted Rand Index uses the gold/predicted contingency table and chance correction. Homogeneity is 1 - H(gold | predicted) / H(gold); completeness is 1 - H(predicted | gold) / H(predicted); V-measure is their harmonic mean; normalized mutual information uses the arithmetic-mean entropy normalization MI / ((H(gold) + H(predicted)) / 2). These partition-only metrics apply to PRIMARY_OUTPUT; candidate-relation rows show n/a.

Provenance

  • Authoritative source: docs/benchmarks/data/stemming-quality.csv
  • Source SHA-256: 5a93a6ab60e46489737cd649eb1ac48182114b9038f7f20195ab9d1c1fc0dd28
  • Evaluation command: ./gradlew stemmingQuality
  • Dictionary language: RU_RU
  • Processing modes: ALL_WORDS, LOWERCASE_GROUPS_ONLY
  • Stemmer versions and transitive artifacts: resolved by the repository's JMH Gradle configuration and gradle.lockfile
  • Radixor version, Git revision, generation date, JDK version, operating system, and dictionary revision: not recorded in the authoritative CSV
Reference in New Issue View Git Blame Copy Permalink