The more I think about it, the more I think having multiple editions on the shelves (digital or physical, whatever) is perfectly valid. I don't think simply reprinting the books is the answer, though. Something more like 5e, or the blog author's example, where redundant information in minimized. The kicker is what edition do new supplements support? Is it possible for a supplement to support all editions? That still sounds like 5e, actually.
If it were me, this is what I would do.
1. I would continue work on 5e, and it would be main edition I would plan for WotC to support and produce material for. Even if people don't end up adopting it, it will still create buzz and web traffic because it's new.
2. I would make the whole back library available as PoD and PDF. The PDFs are out there on the web anyway, so even if people keep pirating them at least WotC could get a slice of that pie rather than just ignoring it in a huff. PoD offers something that piracy can't, actually books. There are a lot of people that like reading books a lot more than PDFs. And, I think the ability to buy products in new condition would be a big plus for many people. I would roll out the back library a few products a month to maximize sales, rather than drown customers in a glut.
3. I would keep DDI as a fairly inexpensive subscription service, however I would shift the focus from gaming aid to content delivery. I wouldn't ditch the former, but it wouldn't be my main selling point anymore. WotC would produce articles, adventures, monsters for 5e and make it available to subscribers, this would be a proving ground, things could be altered based on feedback and then the best of the best would move on to published products.
3b. The real trick would be this, I would add a ratings system to DDI, let users flag what they think is the best material. Then I would use this system to support editions past. Rather than wasting man hours and resources on chasing a smaller market, I would let the fans do it for me. I would invite users to produce material for their editions of choice, and then I would use the ratings system to let the cream float to the top. Top rated material would get its authors rewarded and be formatted for PoD with a good percentage of the sales going to the authors. I'm worried less about making money on sales of a niche product to a niche market than keeping people subscribing to DDI so they can enjoy continuing support for the edition of choice. I want all those retro clone authors working for me, and directing them to my service, rather than pulling people away. I think a lot of them would be pretty enthusiastic to get actually write for the "real" Dungeons & Dragons.
I think if it were done correctly DDI could foster an incredible sense of community. Get people to cross over various editions based on high rated products getting good buzz and reviews (I would want a review system too, and top rated reviewers should be rewarded with small rewards like free subscription or swag or whatever).