Re: [boost] [gsoc] Suggestions on Proposal for Boost Document Library Project

18 Mar 2015

      On 18/03/15 12:18 PM, Antony Polukhin wrote:
...
2015-03-18 19:07 GMT+04:00 Stefan Seefeld :
<...>
...
Second, I don't think an interface to existing office suites is the
right approach to the problem. Rather, I would suggest something based
on existing standard technologies such as XML (and DocBook in
particular), to support the manipulation of structured documents.
Without using existing Office suit API, student will be forced to rewrite
the functionality of Open Office from scratch. Working with spreadsheets is
not just parsing document, but also evaluating functions, plotting charts
and so on... This is hell and nightmare. And it will take insane amount of
time ( about 2,190 years of effort
https://www.openhub.net/p/libreoffice/estimated_cost)
So the approach with unification of APIs seems right to me.
Well, yes, that's what I meant with "manipulating documents on a
semantic level". I agree, writing this from scratch is wrong. But just
providing a programmatic interface to an office suite seems ill-designed
to me, at least for a project other than LibreOffice itself.

For Boost I think one should at least attempt to build the functionality
on top of a standard document model such as XML/DocBook.

        Stefan

-- 

      ...ich hab' noch einen Koffer in Berlin...