processing txt files

[eshellman]

Online Ebookmaker is set up to choose a single file for processing. But most PG submissions include a txt file, because Ebookmaker prefers to use txt source files to make the txt file presented to users (these include boilerplate headers and footers, and are uniformly normal-form UTF8 files).

As a result, txt files are typically not processed and checked by Online Ebookmaker.

Should online ebookmaker process txt files in the upload zip in addition to an html files it finds? This would add complexity to the script. Or should it present a user option to process a txt file from the uploaded zip?

[windymilla]

If txt source files are sometimes uploaded to PG containing "errors" that would be corrected by ebookmaker, e.g. not normal-form UTF8 files, then it would seem sensible for there to be a way for uploaders to check their text file along with their HTML.
It doesn't sound like a user option is necessary. Typically the zip contains an HTML file and some image files. If the zip also contained a txt file, that could be processed/checked by online ebookmaker.

[eshellman]

Ebookmaker processes txt files 99.8% correctly into utf8. that leaves 150 bad txt file in the corpus, which the errata team has been fixing. The question is whether Online Ebookmaker should try to handle cases where html and txt files are uploaded, or just to make it clearer to users that it has not attempted to check the accompanying txt file. I'm leaning towards doing the latter (which I can do) and add txt->txt processing if there is demand for it.

[windymilla]

In that case, with only a few problem txt files, leaving txt->txt processing until there is a demand sounds like the best option.