I don’t understand that. It is not “a better way to work”, it is a better way for *some people* and it is worse, or even impossible for other people. Some people already have TWAIN-based applications, and if the scanner cannot provide images through TWAIN, it cannot be used with those applications.
Barcode recognition, OCR, and PDF/A are useless if you are trying to feed documents into a document management system that expects to use its own barcode recognition and OCR, and to generate PDF/A. If you have a medical records system with TWAIN support, what good does it do the end-user to scan into a folder on a network? They have to go back to their workstation and do some kind of ‘import’ operation, which is rarely as simple as clicking a ‘scan and attach’ button, if it even can do the same thing. What quality of OCR do you get from a 150$ Brother MFC?
Who wants to program their MFD for their individual needs, when with other scanners they just click a button in their application?
I understand completely that the scan-to-folder process is good for many purposes. But it is not a replacement for TWAIN support, it is complementary: Some people need one, some people need the other, some people need both. Neither replaces the other in meeting the needs of the marketplace.
And what on Earth is the point of doing a bad TWAIN driver? Creating ANY TWAIN driver is a lot of work – assigning an engineer or hiring a consultant, working out the user interface, parsing all the triplets, packing and unpacking capabilities, coding the UI, tracing and debugging, doing QA, creating a manual, etc. etc. – A bad TWAIN driver costs just as much as a good one! It is stupid to go to the cost and trouble of creating a TWAIN driver, without making it work correctly! Why not just go out in the parking lot and set some money on fire?