... it shows me what a void there is still to fill!
ISI has had a fantastic run providing bibliometric research
tools
for
nearly 30 years, but only to deep-pocket libraries. WebOfScience
finally brought the ISI analysis of academic pubs into 21st century,
and so it is no surprise that they quickly bumped into Google, who
brought fundamentally the same insight (citation impact/in-degree is a
great clue) to the Web. If GoogleScholar has simply nudged
Thomson (who bought ISI in 1992) to broaden the market for this tool,
that's already progress in my book.
For now, the interesting part to me
is a compare/contrast of just what
each brings to the party. While this review
by Péter Jacsó' (his
earlier review is also helpful) is part of Thomson/Gale's site, I
think it's unfair to see it simply as a vendor whitepaper; he
identifies serious flaws in GoogleScholar. But even with the
price differential aside, it must be clear to all that WoS has some
serious issues, too!
(Some of you might be interested in an author-focused comparison I did
recently between GS and WoS: Scientific
impact quantity and quality:
Analysis of two sources of bibliographic data
, arXiv.org preprint arXiv:cs.IR/0504046, 11 Apr 05). Do
they really want to hold up the interface
to WoS as a virtue?! Checkout the touchgraph browser for
CiteSeer as an example of what we can hope for.
And while there isn't yet an API to GoogleScholar, screen-scraping at
least lets us do some experiments over this corpus; WoS does not
seem willing to provide similar access (I've tried:).
These aren't the only two vendors, of course: GoogleScholar
was certainly inspired by the CiteSeer
(originally at NEC, now at UPenn) project; it continues to be an
innovative force. Our local, generally well-stocked library
doesn't carry Scopus (too expensive?), but I hear good things about
it. Entrez/PubMed
has been mentioned and (while it is great in many other dimensions!) I
don't see it is as especially relevant until the citation linkages it is beginning
to build via PubMedCentral
come online. And when the NIH's "Open Access" policy (cf. [Science
11 February 2005; 307: 825 DOI: 10.1126/science.307.5711.825],
but not without a subscription:) starts to kick in, and as changing
standards regarding exchange of ``open citation'' information (e.g, CrossRef) propagate, the pace of
change is bound to accelerate.
Looking a bit farther afield for suggestions of what might be
coming,
some of you lawyer-types may appreciate what Shepards does for case
law searching. They orignally started doing simply the manual
"inversion" of citation links that ISI does, but grew into an entirely
new source of independent analysis of the arguments connecting the two
documents. Imagine how
helpful it could be if scientific and web citations carried as much
third-party (ie, from neither the cited or citing authors) metadata!
ok, so there is lots to say about authoring by a single individual vs. wiki-mediated hordes (and i suppose James Michener's enterprise falls somewhere in between?:).
but what does the slashDot crowd make of Lessig's ARGUMENTS in Code, about code? what should CodeV2 make sure to address? are you going to weigh in?
Slashdot Mirror
... it shows me what a void there is still to fill!
ISI has had a fantastic run providing bibliometric research tools for nearly 30 years, but only to deep-pocket libraries. WebOfScience finally brought the ISI analysis of academic pubs into 21st century, and so it is no surprise that they quickly bumped into Google, who brought fundamentally the same insight (citation impact/in-degree is a great clue) to the Web. If GoogleScholar has simply nudged Thomson (who bought ISI in 1992) to broaden the market for this tool, that's already progress in my book.
For now, the interesting part to me is a compare/contrast of just what each brings to the party. While this review by Péter Jacsó' (his earlier review is also helpful) is part of Thomson/Gale's site, I think it's unfair to see it simply as a vendor whitepaper; he identifies serious flaws in GoogleScholar. But even with the price differential aside, it must be clear to all that WoS has some serious issues, too! (Some of you might be interested in an author-focused comparison I did recently between GS and WoS: Scientific impact quantity and quality: Analysis of two sources of bibliographic data , arXiv.org preprint arXiv:cs.IR/0504046, 11 Apr 05). Do they really want to hold up the interface to WoS as a virtue?! Checkout the touchgraph browser for CiteSeer as an example of what we can hope for. And while there isn't yet an API to GoogleScholar, screen-scraping at least lets us do some experiments over this corpus; WoS does not seem willing to provide similar access (I've tried:).
These aren't the only two vendors, of course: GoogleScholar was certainly inspired by the CiteSeer (originally at NEC, now at UPenn) project; it continues to be an innovative force. Our local, generally well-stocked library doesn't carry Scopus (too expensive?), but I hear good things about it. Entrez/PubMed has been mentioned and (while it is great in many other dimensions!) I don't see it is as especially relevant until the citation linkages it is beginning to build via PubMedCentral come online. And when the NIH's "Open Access" policy (cf. [Science 11 February 2005; 307: 825 DOI: 10.1126/science.307.5711.825], but not without a subscription:) starts to kick in, and as changing standards regarding exchange of ``open citation'' information (e.g, CrossRef) propagate, the pace of change is bound to accelerate.
Looking a bit farther afield for suggestions of what might be coming, some of you lawyer-types may appreciate what Shepards does for case law searching. They orignally started doing simply the manual "inversion" of citation links that ISI does, but grew into an entirely new source of independent analysis of the arguments connecting the two documents. Imagine how helpful it could be if scientific and web citations carried as much third-party (ie, from neither the cited or citing authors) metadata!
ok, so there is lots to say about authoring by a single individual vs. wiki-mediated hordes (and i suppose James Michener's enterprise falls somewhere in between?:).
but what does the slashDot crowd make of Lessig's ARGUMENTS in Code, about code? what should CodeV2 make sure to address? are you going to weigh in?