* Posts by ~ge~

1 publicly visible post • joined 16 Jun 2019

Large Redmond Collider: CERN reveals plan to shift from Microsoft to open-source code after tenfold license fee hike

~ge~

Re: If they were looking for customers to move to Azure...

Exactly! The problem of legacy MS Word specific documents and the many workflows they fuel is compounded by volumes of new native documents locked into Microsoft applications. Because of this application lock on corporate and organizational document information, Microsoft can charge whatever they want.

That said, we have notified CERN that we can unlock these native documents and open them in an HTML5 browser in a process designed for high fidelity viewing and collaborative editing. The process was also designed for a complete round-tripping of these documents back into existing workflows; without breaking the documents or the workflow they fuel.

The layout engine we developed is local, running inside a browser using WASM. Also runs in a Lamda container. The thing is though, this engine was developed for feeding native documents into Artificial Intelligence machines for analysis; and then merging the vectored results back into the original style. It is designed to enable AI machines to markup, comment and collaboratively edit, interacting directly with the parties responsible for those documents.

This is very cool stuff but with some current limits. For instance, formatting in the browser is currently limited to picking up the existing format. Since there are billions of these native documents already in existence our objective was that of unlocking the information for use in a multi-application environment. We are not much interested in "creating" new documents. Maybe someday. But at the moment the need is to feed AI,

The browser layout engine was also designed to work with document management systems where specific AI machines could be selected and engaged as part of the one-click opening of the document. Which is consistent with our objective; enable many new applications to access volumes of information locked into previously application specific documents. The API for this is currently a work in progress but looking great.

It might interest some to know that at the first face-to-face meeting of the original OASIS ODF TC there was unanimous agreement on a few things. We split wide open on the more important issue of interoperability versus high fidelity compatibility with Microsoft application specific documents. Interop won that argument.

The TC however was unanimous in the opinion that for the world to move forward, many new applications would need to access our knowledge base of data, documents, and methods.

I think that is still the case. The world will insist that we take our information with us into the future. Value added will always trump costly and disruptive rip-out-and-replace.

~ge~