As we have planned earlier about the improvements, the first thing was to go through
1) Referential docs
- OT rules specifications for bengali script by microsoft
- Unicode core specification documentation
- Following AGL specification (We are just following syntax and not naming stuff yet. It will happen with script improvement for generating sfd to ttf)
- Bengali font have lot many conjunct ligatures, since we are following AGL syntax, for glyph-naming step, they are exceeding 31 character limit. For such cases we are removing the script tag placed at the end(i.e.beng)
- though we had script written for these glyphs namings, it would be generated afterwards
- for this, we are presently following the microsoft OT rules specifications for bengali script 
- Previously, with lohit version 2.5.3, there was no lookup defined for akhand forms, so that one is added in newer version
- One more thing that has to be updated to Below-base form substitution is, previously used ligature is different than what it has to be.
- Amongst these two in above screenshot, the rules has been written for the one on the left, but as per unicode specifications & MS typography specifications, it has to be the one on right. so it has been updated now .
- so with this, we are done lookup writings for basic shaping forms. now we are at mandatory shaping forms
apart from this, the things yet to complete are :
- To complete lookup writings for all of the forms
- Removal of glyphs which no longer needed
- Test-File creation
- bug reporting, solving, & reviewing
So that's it for bengali right now. the development is been tracked at lohit2/bengali . surely try to complete the other things asap. For any comments or issues github link is here .