XSweet: New Approach
XSweet does both normalization as well as font based style identification. We need to separate both these services as independent services.
- Normalization (Clean-up, font normalization, junk characters, etc.)
- Style identification
Because of this synchronized activity, we are getting issues like block inside block, and other issues that are unnoticed as of now:
Example:-
....{{some content}}
This results in non-rendering of math tags in the editor due to block scoped schema.
We also find the following issues in Xsweet in terms of LaTex:
- math-display comes inside table cell
example :-
\mu
- Missing closing tags for \left:-
Xsweet latex: \left(7\right)p\left(1\right)=\theta _{y}\varphi \left(\frac{1-\mu }{\mu }
Corrected latex: - \left(7\right)p\left(1\right) = \theta_{y} \varphi\left(\frac{1-\mu}{\mu}\right)
-
Unwanted closing tags:-
Xsweet latex:- \dot {\iota }=\beta _{\iota }\cdot (\iota ^{T}-\iota \right)
Corrected latex:- \dot {\iota }=\beta _{\iota }\cdot (\iota ^{T}-\iota)
-
Fraction correction:- Xsweet latex:- \sfrac{1}{2}