Category Archives: JURN metrics

How is the JURN ‘headline’ total calculated?

03 Monday Jan 2011

Posted by in JURN metrics, My general observations

≈ 1 Comment

How is JURN’s total number of ejournals calculated? Well, it’s a number that’s been built up incrementally since the beta. But a rough checking calculation at Jan 2011 would go something like this:—

* 2,580 non-duplicate English-language titles, as listed in the 2,700-link JURN Directory.
* Around 620 marginal ‘partly in English’ titles that I didn’t enter in the English-language JURN Directory, and yet which are not on Revues / Persee / Dialnet / Cairn since they’re not in Spanish/French.
* 2,792 Spanish titles in the arts and humanities and ‘philologies’, as currently indexed via Dialnet. 1,367 titles on Dialnet currently have full-text, and perhaps 600 of these offer articles relevant to the arts and humanities. *
* 105 older French titles on Persee.
* 274 titles on Revues, perhaps 260 if a handful of geography titles and ‘collections’ are discounted.
* Around 250 relevant titles via the Hungarian central index.
* Around 180 relevant French titles via Cairn.
* Around 250 humanities titles from the Central America region via Redalyc.
* Around 170 arts and humanities titles in Portuguese, via Livre.
* Perhaps another 200 (perhaps more) non-English titles from various national amalgamation services such as those in Serbia, Taiwan, Singapore, Mexico, Catalonia, etc.

That would give a grand total of about 5,200 titles indexed.

However, if only those titles that carry at least some English articles are to be counted, then the calculation is more like: 2,580 English titles + 620 known partly-English + another 700 unknown partly-English titles hidden among the total at Dialnet / Redalyc, et al = 3,900.

* Indexing Dialnet does bring in some article record pages that don’t contain links to full-text. JURN indexes Dialnet via three URLs that bring in: i) just the main index pages for journals (not the TOC pages), ii) full-text PDF articles hosted on Dialnet, and iii) article record pages. The latter may or may not contain links to full-text (I estimate about a one-in-six chance of full-text from a Dialnet record, in arts and humanities searches). However, for those searching for English search-terms, this is unlikely to flood the search results with masses of records that only contain citations / abstracts. I think it’s a price worth paying, when weighed against the wealth of full-text material that it can bring in for a searcher.

JURN by country

25 Thursday Jun 2009

Posted by in JURN metrics

≈ 1 Comment

The table below shows each nation’s proportion of site-index URLs in JURN. It’s not a table of visitors to the front-page.

To get it I ran the JURN site-index backup XML through an URL extractor and then Microsoft Word (replace all “/” with “,” to get a comma-delimited list), then Excel so I could delete a few columns, and then output a plain text list of circa 4,000 root hostnames.

I then loaded this plain hostname list into the excellent little Windows application Country Whois (which seems to be fully-functional trialware, and which based on a few of my tests uses a more precise lookup database than lesser freeware), to output some basic geolocation statistics by nation…

jurn-location

It’s not too precise — the software seems to have de-duplicated and thus removed hundreds of URLs, whereas the original list would have shown perhaps three or more ejournals hanging off the same university URL. Keeping these in would probably have boosted the results for the U.S., the U.K., and Australia even further, because that’s where the bulk of multi-journal university websites are located.

And the list also strongly underestimates the number of journals indexed from a couple of countries, where lots of journals can be brought into JURN by adding a single URL. I estimate that the following numbers should thus be added to the above table…

France: add Persee-hosted (30) + Revues-hosted (about 140) = 180.

Spanish: Dialnet- hosted (434, with perhaps ¼ arts & humanities) = 108

And perhaps around 15 each, added to Hungary, Catalonia, Russia, Mexico, Taiwan, and Brazil.

I daresay if Japan made it easy to add all their journals (in English or not) via one URL, then their total would be boosted by 100 or so. And a one URL solution for Germany could have added another 100 or so titles. So far as I know there is no such “one URL” option for Japan or Germany, other than in science.

But China, where is China? A mere two URLs that are not based in Taiwan or Hong Kong?

JURN Directory checked and refreshed

15 Monday Jun 2009

Posted by in JURN metrics

≈ Leave a comment

I ran the excellent Linkbot v5 link-checking software over the JURN Directory. The results are: five dead links removed (also removed one open journal that has sold out and is now located behind a paywall at SAGE); 15 moved links re-located and repaired; and I also removed a couple of links to journals that had only marginal free content.

If you are keeping a local saved copy of the JURN Directory page on your desktop (possibly to avoid the “bouncy puppy” concertina effects, which vanish if the page is saved locally), please refresh it.

Getting only the free articles into JURN

13 Saturday Jun 2009

Posted by in JURN metrics

≈ Leave a comment

Someone asked about what comes into the JURN index, when a title is indexed but only offers a limited amount of free full-text or “free-sample” articles. Does the rest of the online material (link-less tables-of-contents, abstracts with no full-text links etc) from the journal also enter JURN? The answer is: no, not usually. It’s usually possible to filter at the URL level so that only the free content enters JURN. For example, by only indexing URLS such as:

http://www.journal.com/journal/sample/*.pdf

http://www.journal.edu/journalABC/documents/*.pdf

A real-world example is:

http://www.egyptpro.sci.waseda.ac.jp/pdf*/*/*.pdf

Where “*” is the Google CSE wildcard. Of course if some dimwit IT techie then decides to juggle the directory structure, it will erase the journal from JURN. But that’s a risk any directory or search-engine takes.

Sometimes a few PDFs to do with society or journal administration matters can be called into search along with the articles, if all the PDFs sit indiscriminately in a single URL path. A search for:

site:http://www.scholarly-society-journal.info/ filetype:pdf

… will usually show if there are too many of these. Google tends to bunch that sort of material at the top of site: search results. Usually there are only a dozen or so.

It’s different with the few ejournals that cheekily use standard ‘open access’ publishing software, but which actually keep recent articles locked away behind a one-year or even three-year rolling paywall. The software is not intelligent enough to place paywall article abstract pages on a different and distinctive URL path, and then to automatically transfer&bounce these when the article becomes free. But by indexing only the .pdf path in such cases, that will usually call only fulltext articles into JURN.

An academic search group-test

11 Thursday Jun 2009

Posted by in Academic search, JURN metrics

≈ 4 Comments

I decided to do a quick group-test of search-engines, based on an unsophisticated casual academic search for the keywords:

Mongolian folk song

I was trying to mimic what might be a typical student search. This is what I found that was free:

The main Google index:

Top result is a dubious spammy-looking link that actually leads to a clean webpage for the commercial album Mongolian Folk Songs, with embedded audio clips for each track. Nice. The next two links are YouTube videos. At the foot of the first page is a link of some use, a short English-language 2005 Xinhua press-agency story carried by the Chinese People’s Daily newspaper talking of “1100 Mongolian folk songs rescued”.

Microsoft Bing:

Not bad, not bad at all. Better than Google by far. The Wikipedia page “Music of Mongolia” is result number one, but result number 4 gets top marks – a direct link into the Mongolian-run UNESCO-accredited International Institute for the Study of Nomadic Civilizations website, detailing (in good English) a major fieldwork song-gathering project now underway, “Heritage of the Mongolian Long Folk Song” (2008-2010). Result number six is also strong – a link into the new Smithsonian Folkways magazine, recommending two albums.

Google Scholar:

Oh dear. The top result is free. But it’s a PDF of the vintage book The Souls of Black Folk (1903) by W.E.B. Du Bois, and this only seems to mention the word ‘Mongolian’ in passing. Misleadingly, the date of the book is labelled in the results as “2007”. There’s only one other free result on the page, a Google Books link to Mongolian music, dance, & oral narrative: performing diverse identities (University of Washington Press, 2001). Tucked away at the bottom of page three of the results is some free and useful full-text, the 1997 article “Mongolian Oral Epic Poetry: An Overview” from the scholarly journal Oral Tradition. Everything else on the first few pages is trapped behind a paywall.

Google Book Search:

Mongolian music, dance, & oral narrative: performing diverse identities‎ appears again, and is number one. The link for it gives a “limited preview” link that leads to a deep interior page discussing the “overhaul” of classification of different types of song under communist rule. There’s only one other “limited-preview” result on the first page, linking to the book The Horse-head Fiddle and the Cosmopolitan Reimagination of Mongolia. On page two there’s another “limited-preview” book, and it’s the 2008 Lonely Planet guide-book to Mongolia.

Google news:

A surprise. Not bad, if you’re looking for contemporary performance. Number two is a Financial Times arts report “Steppes of Dreamers, Ukrainian Pavillion, Venice [biennale]” and talks of… ‘a Mongolian folk song, deconstructed into three elements of wind instrument, voice and whistle’. Then there are various journalistic puffs for folk performances. Halfway down the page is a PSFK trend-spotting agency report on a free 50-minute “Podcast Documentary on Chinagrass – Contemporary Chinese Folk Music, performed by Hanggai … a Beijing-based Mongolian folk band composed of 5 members who feature the distinctive Morin Khuur and throat-singing.” Sounds great! The article is Creative Commons, too. Near the bottom of the results is a link to China Central Television proclaiming “Xinjiang preserves ethnic folk arts”.

Bing news search:

“We did not find any results for Mongolian folk song.” Some way to go, I think.

Findarticles.com (free):

The first four results are duplicates of a short 2005 press story “China, Mongolia to protect endangered ethnic song”, then the results default to articles on other unrelated types of folk song.

Intute Arts and Humanities:

Zero results. If the search terms are changed to Mongolia + song then I get a record for the “Music of Tuva” website. Tuva being near to Mongolia, and now in Russia.

Archive.org texts:

Zero results. If the search is limited to just ‘Mongolian’ then I get a wide range of books including an avalanche of dusty pre-1920s linguistic studies, and er… “Racial Origins of the Jews – Eugen Fischer. An article from the defunct neo-nazi magazine”. Oh dear. In amongst the avalanche there are several scanned editions of Sagas from the Far East: Or, Kalmouk and Mongolian Traditionary Tales (1873).

OAIster:

Zero results.

Intute UK repository search:

Zero results.

DOAJ search:

Zero results.

Open J-Gate:

Zero results.

IngentaConnect:

Zero results.

JISC ticTOCs: (search tables of contents from major journal publishers)

Zero results for a search in titles or subject. Even the word ‘Mongolian’ on its own found no results.

CiteSeer:

Mostly science, but I thought there might be some linguistic or ethnographic materials indexed. I used the search: Mongolian AND folk AND song, and included citations in the search. Two results, neither relevant. Using Mongolian AND song obtained more, but not better, results.

Sweet Search

Supposedly an academic search engine, the top results were from www.emusic.com (trying to sell me MP3s), the state-owned www.chinadaily.com.cn, www.npr.org, and bbc.co.uk. To be fair, it does a good job of clearing the web of spam, but the lack of academic articles in the results shows that it’s aimed at school children rather than those at university.

Scientific Commons:

One result, to what is now a dead ‘404’ link.

EBSCO Open Science Directory

Zero results.

China Academic Journals full-text:

Just three results from a 1915-2009 search, one in English (“Mongolian Folk Song and Dance Troupe Visits China”, which was a short news report from the state-run Voice of Friendship magazine).

Scirius:

The first two results are useful, but both lead to “404 not found” messages. The first page of results show that Scirius search is confused by Chinese science authors whose surname is “Song”, and by references to the Song dynasty.

The British Library: (“search 9 million articles from 20,000 journals”)

Zero results. Did someone forget to plug the database cable in?

HathiTrust 0.2 beta

15 results for full-text items when the search was limited to: “mongolian folk” song. No result was relevant, and the results included eight instances of hits from ‘Library of Congress subject headings’ lists.

Journal TOCs

One result, an article in the commercial paywall journal Acta Orientalia, “Dsakhchin (West-Mongolian) folksongs with Buddhist content”.

BASE:

Seven results. Number one was in Hungarian, and was a description and track-list of a Hungarian world music CD. Number two was (oh dear!) our old friend Du Bois, W.E.B., The Souls of Black Folk (1903). Number three was a full-text PDF Thesis titled “What is Throat Singing?”, which is the result of fieldwork in southern Siberia. Number four was an English description of the Czech book Kazakh folksongs : from the two ends of the steppe (2001). The rest of the results were junk from Encyclopedia of World History (a spurious result, which anyway bounced to the Bartleby.com front-page), and another spurious result from an atrociously bad OCR copy of the Deseret News newspaper from 1878.

JURN:

On the first page, the full-text of:—

1. IIAS Newsletter. A long review of a 2005 CHIME Foundation conference which asked “Do performers of ritual music in East Asia address their performances primarily to the gods or to mortals?” starstarstar

2. Asian Folklore Studies. A review of two books from the late 1980s (On Huaer and Selections of Traditional Qinghai Folk Songs), which only mentions Mongolian songs in passing. star

3. Oral Tradition. “A Comparative Study of the Singing Styles of Mongolian and Tibetan Geser/Gesar Artists”. starstarstar

4. Asian Folklore Studies. English reviews of two German books on Mongol epics and epic songs. starstar

5. Ethnomusicology OnLine. A review of the commercial CD Mongolia, Living Music of the Steppes: Instrumental Music and Song of Mongolia, with three sample tracks as embedded audio. starstarstar

6 and 7. China Heritage Quarterly. Two articles, the somewhat-tangental “Cultural Heritage Properties of Qinghai, Gansu and Ningxia: Performance Items” and more usefully “A Tale of Two Lists: An Examination of the New Lists of Intangible Cultural Properties” (a very long account of the history of Chinese attempts to preserve folk cultures and later UNESCO involvement). starstarstar

8. Echo. (A tangental result in a long article about Nepali pop music, due to the titling of an album as Mongolian Heart). star

9. Oral Tradition. “Mongolian Oral Epic Poetry: An Overview.” starstarstarstar

10. Asian Folklore Studies. A fair but critical scholarly book review of Mongolian Music, Dance, and Oral Narrative: Performing Diverse Identities (2001). starstarstar

Highlights on further results pages include: “The Mechanisms of Epic Plot and the Mongolian Geseriad”; “Mongol creation stories”; “Teaching of the Silk-Road Epics: a workshop in Turku”; “Folk Ecology and Rural Epics in China” — and all found without focussing the search-terms or using any Google search modifiers.

And if the Graduate Journal of Asia-Pacific Studies didn’t use stupidly-huge dynamic scripted URLs (all hanging directly off the main university URL, durh) I daresay that the excellent full-text article “Blue Heaven, Parched Land: Mongolian Folk Song and the Chinese State” would also show up in JURN. A researcher could get to it at the hosting university via a specific Google article-title search.

My library catalogue (inc. the Birmingham Conservatoire music library):

No results on a keyword search. However, a direct title search finds one copy of the book Mongolian music, dance, & oral narrative : performing diverse identities (2001). But it would be cheaper for me to buy it on Amazon, than to pay for a train ticket to specially go and get it.

Project MUSE:

The first four results are strong (although not free full-text to the public), but then the results turn to mush — and by the bottom of the page we’re back to… W.E.B. Du Bois and his The Souls of Black Folk (1903). Very worthy, a seminal thinker, and all that… but not relevant to the search.

JSTOR:

JSTOR coverage is strong (340 results, inc. articles from the back-issues of Asian Music, Journal of the International Folk Music Council, British Journal of Ethnomusicology/Ethnomusicology Forum, Harvard Journal of Asiatic Studies, Journal of Asian Studies, and Far Eastern Quarterly) — but you’ll only obtain them if you or your organisation have access to JSTOR.

And if you’re lucky enough to have a university that subscribes to RILM Abstracts of Music Literature and Music Index Online, you should be assured of a decent starting bibliography — even if accessing the full-text might still prove tricky.

Ole!

09 Tuesday Jun 2009

Posted by in JURN metrics, New titles added to JURN

≈ Leave a comment

Nice; over 1,100 Spanish-language open access ejournals, in an elegant and uniform combined directory/TOCs/article-server format. No registration required, no silly fuss about “on-campus access only”.

And gasp in awe — it has uniform-format article-linked tables-of-contents across more than 1,100 titles. Gasp again when I say it seems to be up-to-date, PDFs speed to your browser faster than a hungry ferret, and it has no broken links I could find.

Either the Spanish set up a wonderful infrastructure circa 1999, or they’re archiving by hand?

And before you ask; yes, the articles it offers (mirrors?) are showing up in Google in a managable way — and are thus now also experimentally showing up in the JURN search results. But I won’t be adding another 1,100 to the JURN home-page title total.

It might have been better if they had set a distinctive URL-string for the open arts and humanities titles, though, rather than having all types of journals sitting on the same URL-string. This means the articles are included in JURN at the price of bringing in some pages that just give a basic reference only, and not all articles are from arts and humanities titles (though around 2,000 are, both pay and open) — but that shouldn’t trouble most people searching in English for arts and humanities phrases. I think it’s a price worth paying, just this once, for such a huge ‘one-URL’ haul of full-text articles. Many thousands of which are in English, by the way.

A weed among the shemales

01 Monday Jun 2009

Posted by in JURN metrics

≈ Leave a comment

I’ve weeded some crud from the index, a result of running a range of test searches with JURN. The following were removed:- six publication titles (spam infestation, mostly in trade and fashion magazines); ten poor-quality blogs (they’d set up shop within various journal websites); and two ‘artist directories’ (again, set up within magazine websites).

JURN — now probably the only search-engine on the planet where you can safely search for shemales, viagra and similar, and actually get useful results. Heh.

URL fragments

06 Friday Feb 2009

Posted by in JURN metrics

≈ Leave a comment

Someone enquired if I was aware that Google Custom Search “Sites” control panel needs to have a distinctive fragment of an URL entered if it’s to seek accurately for a match. This is to do with checking if an URL is already in the site index. Google Custom Search will happily add a duplicate if you let it, especially if you just check against a full URL. Yes, I’m fully aware of this feature, and that it’s case-sensitive — and so only search for a distinctive fragment of an URL, before adding it to the site index.

I am also cutting back the URL slightly if needed, to give Google a slightly wider “spread” in what it picks up from that URL. Ideally, I’m seeking out just the URL that holds the articles (which is often very different from the main journal URL).