1. PDF Issues

1.1 Distiller vs. PDFWriter

Distiller creates PDF much more reliably than PDFWriter, because Distiller is a two-pass process, while PDFWriter is merely a one-pass process. Unfortunately, this complicates PDF construction for students who are novice computer users. To further complicate matters, some printer drivers convert color into black and white.

We request that Adobe create application an plug-in that implements "Export to PDF..." at the menu-item level (for Word, Word Perfect, etc.). The print metaphor has too many mismatches.

2. Library Issues

2.1 ps2ascii

Program ps2ascii converts PDF (and PostScript) into ASCII. This is suitable for some full-text searching. Coupled with OmniMark, we could create SGML versions of documents in the online collection. We have no resources for this endeavor, however. Should we try to plan these conversions now?

2.2 Submission form

File uploading continues to be awkward for students, especially when they must resubmit multiple files. Java applets cannot read local hard drives, and therefore is not viable, nor we do not want to support full-fledged client-side applications. We are torn between the maintenance issues of server accounts, training users to use ZIP+FTP, or continuing to use the same klunky upload system. How feasible is an SGML-template email with attachments? Can we settle on TAR-GZ or ZIP for submission of multiple files? (We must support Mac users.)

2.3 Very large files

What is the largest believable size for a single submitted file? Issues: upload, download, backups, portability. Today? In 5 years?

3. ETD-ML

3.1 ETD-ML, V1.1

We are considering reworking ETD-ML and releasing V1.1.

Although we have not decided how to proceed, we have received suggestions as follows:

• Change the organization of ETD-ML documents so that each chapter is a separate mini-document, using HyTime hyperlinks as a "table of contents" to connect the chapters together.

• Include elements <i> and <b> for bold and italic.

• Allow inline images for non-standard characters.

• Allow more control of the presentation (like in HTML), allowing color, background designs, columns, to achieve a more "magazine-like" appearance. Note that this may completely comprimise our current ability to present ETDs in ETD-ML to those with visual impairments.

3.2 Mathematics

ETD-ML v1 has no native math support. Unfortunately, adding it will comprimise the learnability and writability of ETD-ML. How shall we proceed with mathematics (e.g., implement the ISO Maths DTD)?

3.3 DSSSL

We do not have DSSSL scripts to understand ETD-ML and to create HTML and PDF documents from them. Again, we are strapped for resources (i.e., my time is multiply allocated). Is DSSSL conformance critical to our success?

3.4 Inline vector graphics

We need to choose an inline vector graphics standard for ETD-ML. What are the pros and cons of using Encapsulated PostScript (EPS)?