[ Date Index ] [ Thread Index ] [ <= Previous by date / thread ] [ Next by date / thread => ]
On 14/05/11 12:09, Grant Sewell wrote:
On Sat, 14 May 2011 11:54:17 +0100 tom wrote:On 14/05/11 11:29, Grant Sewell wrote:On Sat, 14 May 2011 07:33:57 +0100 tom wrote:On 13/05/11 18:26, Grant Sewell wrote:<snip> *Some* documents are in proprietary formats. Not all. Grant.For proprietary read 'human and not computer'. Its very very easy to take computer interpretable data and present it for humans to read. Its not so easy the other way round. To my PC an ODF document is just garbage compared to the same data in an XML file which can be made human readable with a style sheet. I'd guess than most homes and offices have encrypted (and lost)>90% of their data in documents. Tom te tom te tomFunny, I've just taken a random (fairly small) ODF spreadsheet document on my computer, renamed it unzipped it (since it seems to be a fairly standard zip compressed file), and this is what I found: content.xml: XML document text meta.xml: XML document text mimetype: ASCII text, with no line terminators settings.xml: XML document text styles.xml: XML document text META-INF/manifest.xml: XML document text Thumbnails/thumbnail.png: PNG image, 225 x 256, 8-bit/color RGB, non-interlaced Now, as far as I can tell almost all the above components of that ODF document are XML, which is exactly what you wanted. I'm not entirely sure I understand how you can say "To my PC and ODF document is just garbage compared to the same data in an XML file" when an ODF document comprises almost exclusively XML files. Sure there is a PNG file but I'm not sure normal XML is the best way of storing 225x256 pixels of 8bpp image data. I have no doubt it *could* be done, but is it necessarily the best way of doing so? Grant.No thatâs NOT what I wanted - I want the DATA xml (or JSON or whatever) formatted - ODF is about presentation. When I get an invoice I'd like the DATA in it (dates, address, invoice number, company identification etc) formatted so that my accountancy package can read it straight in.So... what you're telling me is that ODF *doesn't* include my data? It *only* includes formatting information?! What the hell has LibreOffice done with my data?! Where is it?! :p I thought ODF contained both the data *and* the formatting stuff... so *you* could take the data (and just the data) if you wanted, whereas those of us who want the data *and* the formatting information could use it too! Sounds like a win-win to me! Incidentally, if it seems that I'm being a touch facetious it's because I am. I understand completely your viewpoint, and I don't necessarily disagree with it. But the world isn't black/white so surely there's room for 90% grey in your argument? Grant.
No - make that 95% grey....I have spent a long time trying to automatically extract data from documents and its damn near impossible to do it usefully. Even a word document contains your data but if you were to try and extract (say) the address, date and sender of a .DOC letter then your chances of getting it right >50% of the time are pretty low. The point I am trying to make - and I'll admit that its hard for those brought up on documents to see it 'the right way' but if you manage your data then you've got 99% of your work done - and there's no need to make it human readable - the recipient can make it human readable the way they want. So you send me - for an invoice (say) <invoice><who you are/><who I should be/><date/><paydate/><list of items><item 1>....<item n></list of items></invoice> then my computer can read it in a flash, you can style it the way you want, I can ignore the styling of the week and I can take the info I want, the info you wanted to give me, and the paper sized representation that 90% (why always 90%?...) of the last 20 years of computing effort can be ignored and we can advance electronic communications to the level they were 20 years ago but some (rich) idiots thought they'd make a killing selling us a victorian solution to a problem they made themselves.
Tom te tom te tom -- The Mailing List for the Devon & Cornwall LUG http://mailman.dclug.org.uk/listinfo/list FAQ: http://www.dcglug.org.uk/listfaq