Brad Collins wrote:
Brent Gueth writes:
I don't care if XML becomes as common as plaintext and everyone uses it, you can run into a problem in 20 years where XML falls out of favor and there won't be software to render it properly. This will lead poor fools having to redo all the documents all over again. This is not a good thing.
[snip]
As for plaintext, one of the core design goals for XML is that you'll be able to open it in any text editor and read it. If a file is human readable when it's opened in a text editor then it's a type of plain text. All XML does is place tags around text in order to give the text a structure that machines can understand.
Good points. Properly marked up documents, where the XML vocabulary describes the structure and semantics of the text, is highly repurposeable. Should the day come that XML disappears from use, it will be relatively easily to transform such XML documents into whatever is new. Why? As Brad notes it's because an XML document comprises "plain" text which has markup added (the markup itself is also "plain" text) describing what the text is. One can think of markup as simply a sort of descriptive metadata. In the worst case scenario where one can't find anyone to write a script or apply an XML processing application to do the transformation (a scenario which will only happen if world-wide catastrophe strikes), so long as there are running computers with text editors laying around, one can open up the XML document in a text editor, and there is the "plain" text, right in front of you, nicely described with markup. Though it may take some work (depending upon the extent of the markup), and some text metadata information may be lost, one can use the text editor to strip out the markup and restore the content to "traditional" PG plain text -- if so desired. (In essence, XML markup follows Michael Hart's philosophy of using text encoding to digitally preserve public domain Works.) DP plans to apply an intelligently-designed XML vocabulary optimized for book materials to their first-generation masters (they are looking at a well-constrained subset of TEI, such as PGTEI now under development by Marcello and others.) This is a good plan. Jon