DML-thoughts, by Peter Michor

• A statement of the IMU that it endorses the whole enterprise and some ground rules for it (so that editorial boards could relate to it when contacting their publishers or their authors). This should also contain a statement that the IMU is willing to accept copyright for the digitization of material in the DML, and that it gives this into the public domain forever.
• A call of the IMU to editorial boards to contact their publisher about agreements for the digitization of whole runs of journals with a moving wall of 5 years, say. A model copyright statement for this material to reside with: national mathematical societies, the IMU, or the EMS.
• Some discussion on: How does DML relate to JSTOR. Could JSTOR be persuaded to give out the raw data of the digitization freely.
• Coherent applications to the EU and NSF, containing also a description of the management setup.
• A recommendation on format like: use 600dpi, store the initial formal (tiff) by page, and offer a pdf file of the whole article (not page-wise).
• A proposal for the directory structure of DML.
• A beginning list of servers who offer the material. This could induce libraries to start work on it. There were comments on this proposal, and it may be premature. This also relies to a plan how DML should grow: There could be a statement that any institution which contributes can have the whole bulk of the raw material for its own uses (including Publishers), with a clear aim that aftersome time and some level of completion this should be become available to any scientific library in the world to mirror.
• A master file containing all journals and many books, where parties can indicate (by web access) what they want to do: a sort of self organizing coodination of efforts.
• DML is now in use for this enterprise - digital mathematical library. I think that `digital' should not be in the title. What about `World Mathematical Library' WML.