The switch has resulted in broken links: old links to items on gallica do not work any longer, and can't be mended easily/intuitively.  

See Dana F. Sutton's Emergency Note at http://www.philological.bham.ac.uk/bibliography/ :
Quote:


Dana F. Sutton is going on strike because of this. And I myself will not post any links to ggallica items here any longer up to the moment when I can assume once again that their URLs will remain reasonable stable.

En fait, il y a bien des transferts automatiques entre les url de Gallica
1 et ceux de Gallica 2. Mais ce transfert n'a concerné que les liens
utilisant une adresse "ark" de ce type :
http://gallica2.bnf.fr/ark:/12148/bpt6k60572f
et :
http://gallica.bnf.fr/ark:/12148/bpt6k60572f
pour prendre l'exemple de l'ouvrage d'Abati, De admirabili viperae natura
et de mirificis ejusdem facultatibus liber

En effet, depuis plus de deux ans, ce sont ces seules références que la
BnF peut garantir dans la durée, et non les vieilles adresses du type :
http://visualiseur.bnf.fr/Visualiseur?Destination=Gallica&O=NUMM-060572

I received a long and friendly mail from Arnaud Dhermy at BNF. I'll quote the part whis is of probably most interest to those of us who have linked and do link to documents in Gallica/Gallica2:
Quote:

Dear Dr. Dhermy:

This response is not satisfactory. What you call the "old type" http://visualiseur.bnf.fr/Visualiseur?Destination=Gallica&O=NUMM-060572
" in fact rep laced an older URL formula ""http://gallica.bnf.fr/scripts/ConsultationTout.exe?O=n
"). To the best of my knowledge, it was never made clear to users of
your site that URLs of the "ark" type were in any way superior to the
others, and as far as I am aware it was not publicly announced that
"ce sont ces seules références que la BnF peut garantir dans la
durée." At what time, in what public foru m, and to whom was any such
statement made? There seems to be no way by which URLs of the either
of the previous URL types can be converted into URLs of the "ark"
structure by a simple global search-and-replace operation, such
conversion seemingly would have to b e done manually on an i t em-by-
item basis (if I am mistaken about this, I would be grateful if you
would give instructions how an automated conversion can be
accomplished). Therefore in my bibliography it appears that I am
obliged to use URLs of the "Visualiseur" type, because URLs of the
"ConsultationTout" structure can aut omatically be converted only into
this one (I am, obviously, unwilling to change nearly 4000
bibliographical records manually).

It is remarkable that in its comparatively short lifetime the Gallica
site has identified its posted texts by no less than three different
URL structures, especially considering the absence of any adequate
redirection mechanism that embraces them all. I therefore respectfully
suggest that Gallica ought to make a commitment to its readers th at
the validity of URLs of the "Visualiseur" structure will be maintained
in perpetuity.

Yours sincerely,

Dana F. Sutton