Damien Katz

I haven't followed CouchDb as closely as I probably should but one feature that did resonate was the notion of peer to peer replication. This would be killer for a lot of Notes apps, but I suppose converting those apps to something else with CouchDb underneath is a pretty good trick too.

To anyone thinking along these lines this is obviously a pretty cool tool. The one stated limitation sadly is a loss of rich text field fidelity. So my question to you Damien is what are the technical hurdles involved in that sort of conversion? Is there simply no field type in CouchDb that even closely matches the capability of a Notes Richtext field?

In any event. Congrats on the success of CouchDb. It's been amazing to read about it go from concept to reality.

-Kevin

Hi Kevin.

I haven't looked to closely at the project, so I don't know all the issues.

But there are lots of tools to convert rich text to HTML, both built-into Notes and richer 3rd party tools. So that's probably the best way to go about it, though reaching 100% fidelity of RichText to HTML will always be difficult. I'd guess one obstacle is attaching embedding images to the doc and changing the html markup to properly point to those attachments.

Damien,

So what would be the end resulting data structure in CouchDb of a Notes Richtext field with text formatting, embedded images, the odd attachment, and such? Are there even field/data types in CouchDb as there are in Notes et. al., or is everything xml? Would attachments and embedded images be encoded within the xml or alternately extracted as file resources and referenced in html?

Thanks,

Kevin

The rich text field would need to be converted to an HTML attachment, and the images as attachments too. Then, when GETing the html from the CouchDB server over http, the browser will of course automatically GET the images from the same doc. But to do this, you have to make sure the markup that's being generated has it's relative image urls pointed to the right place to load from the doc on the couchdb server.

Optionally, the HTML can be stored as a string in the document, and the browser can use AJAX to load and display it properly.

It's also possible with HTML to embed imagines directly in the markup, by base 64 encoding them. Though I would only recommend that for small images.