This again makes me doubt if we are talking of the same thing using the words “scaling” or “inline-formula” or “image” in the given context.
As I remember it (currently no ‘Word
’ available here) Word
's included version of MathType
always showed an image
by default , and this image was exported to different applications (Writer, web browsers) when opening a Word document with equations. (At least I had to set anew hundreds of formulas when I abandoned Word and moved to StarOffice in about 1999.)
Concerning the term inline-formula: To my understanding this must be plaint text coming with any kind of markup telling the editor (probably/wishable) and an extension/macros (surely) the fact “This is intended to finally be shown graphically as a formula if it goes to a printed version”. The examples contained in the document “nextExampleMathInlineFormulas.odt” already attached to a previous comment of mine were of this kind using the shorthand markup-parentheses $.
(left) and .$
(right). That attachment also contains an actually working very short (30-liner) macro from my toolbox for the conversion from “inline” to “rendered”. (A revese macro would be very simple, but ´was not provided there.)
Anyway: A typesetting application using inline markup for mathematical formulas needs to be “free” in its ways to position (and anchor and align and…) the resulting graphical elements when creating a printable view of the document or must be told explicitly in what way the author wants to get it done in the given case. An example demonstrating this problem and a very raw first approach to its solution is the “large” formula contained in a TextFrame.
I don’t think developers are blind concerning the Math
component (which is “independent” anyway as far as components of a monolithic application can be), and one of the well known ones (@mikekaganski) already posted a comment to this thread. Two reasons may be relevant for the lack of development:
- Missing human resources.
- Doubts in the probably outdated concept of
Math
(specifically concening the ‘OpenSymbol’ approach.
An actual specification, however, should be based on very profound knowlege.