Devonthink and Opera

Long long time Opera user here. I am working out a workflow with devonthink which now has adequate documentation and looks set to become vital for my work. I believe there will be more and more people perhaps of the type that would tend to have been long time opera users, that would want to have devonthink support as a hard requirement for their browser. To push content into devonthink, there are a set of bookmarklets here . I am not having success on 11.53 with these, erratic or no function. I read that 12 looks to be more stable and will try that but I am not entirely optimistic. I would like to point out that I will need this to work not only in 12 but in all future versions of Opera. If you are unfamiliar with devonthink, it would be a good idea to read the few books on its use before trying to make sense of the program as it is a bit intimidating in all its flexibility.

In my view it could be a good move for Opera to take things up a notch and work out something much better than the competition in terms of devonthink support. I do not want to be rude but I am not having great success with the bookmark menu in its not-nested format. It does though make perfect sense that keeping bookmarks as well as downloaded content organized for serious work use should be out of scope for Opera, and Devonthink frankly appears to be entirely unique - it scales beatifully and does the least possible damage in terms of locking in data. Openmeta, as many perhaps are aware, is a hack nothing like officially supported by Apple. All work organizing information based on openmeta can be cancelled out with any minor update of OSX. With any kind of integration, I and lots and ever more lots of other high-end academic users would be completely locked into Opera long term. I mean you could hack something quite useful quickly with a few applescripts in a few hours, but I would rather take a look at what could be accomplished by a real programmer having a talk with devonthink devs. The issue would be to accomplish a very simple pragmatic protocol for exchange of some well formed data, after which it would probably be not complicated to offer the benefits of both programs to each other. I mean, you could make a sensible bridge with conservative design that wouldn't tend to violate major design principles of good software, and see what could be done with what that could give you. I am not making the case for a shotgun wedding and joining at the hip, more like flirtation so to say. I think an engineer at opera spending a week at this could save me several weeks of dire irritating maintenance work on a yearly basis, possibly much more.

Give me a pm or a response for a more thorough opinion, but some obvious feature ideas off the top of my head:

* Bookmarks synced inbetween opera and devonthink. I could just toss them into devonthink in a heap and have them automatically grouped and sorted into hierarchy in batches, replicated to more than one category where appropriate etc. The latter hardly possible in opera at the moment. I would not hit the roof of Operas solution, and I wouldn't have to leave opera to choose bookmarks. This is likely possible even just using applescript today, but would be a main feature that would be good to have properly designed.

* More shortcuttable buttons for things like web2.5 features. If I do a google search, there could be a button or right click option to have the feed of new results from that search added to devonthink. Perhaps even with a suggested category. A single button to save a copy of the page as a pdf with the google translated version of the article/selected text under the main article.

* Save pdf of page to devonthink and classify or autoclassify directly.

* Search both google and devonthink at once, and say remove google hits linking to already existing documents at the press of a button.

* Sensible work on important sites like pubmed, google patents, google books, a few important newspapers etc. Simple text parsing and clutter removal, working out simple kinks with site javascripts and the like. Perhaps identifying authors and providing them to devonthink for metadata use as such. Ask devonthink devs what will work well with their search engine. This could costs practically nothing to make and it would save constrained time by the metric ton for professionals in a way that you would not have many other means of accomplishing. Most people today are drowning in data. Devonthink is the only solution I know of that provide everything across the specter from tossing it in a heap to combining in a sensible and flexible manner both hiearchical and tag based structure. Tags though, just to have it said, is inferior in devonthink over folders in terms of searchability and AI features.

* Generally some pragmatic parsing which shouldn't be very hard for an opera engineer to offer a few suggestions for; I am making a database to I would want the database searches to see the content. Perhaps I could get a pdf where only the content/selected text would be text, and all menus and other stuff would be an image? I could need the image of the page if the content or site were to be taken off the net, to use the text in an academic reference I better have something at least indicating it was written by such and such at the time, it isn't necessarily fantastic to just have a copy paste of the article text.

* Devonthink is not very usable as an rss reader, perhaps there could be benefit to opera integration here.

* Opera as a default reader of devonthink content, obviously.

* There could possibly be some interesting potential for cooperation in terms of Opera being the owner of some very sophisticated server infrastructure with compressed data transfer and impressive parsing, registered users with secure password storage and such. I mean I would easily pay a thousand kroners just for a few simple buttons linking devonthink and opera. If I could say search a database specifically prepared by Kendrick when researching cholesterol something it would simply be priceless. I believe there are quite a number of people that would be confident having such mildly sensitive data on Opera servers that would not in the same manner feel like having that data accessible through google something. In medical research as an example, personal data export prohibition could be a real issue. There are real agreements and treaties on these types of questions inbetween European countries, but a very weak legal framework across the atlantic. More than legal questions though, I believe it is about the idea. Opera would be credible in making something genuinely tailored to academic needs as well as ideas. Technically incredibly simple stuff could go a very long way in terms of features. AI bayesian search without any crud isn't provided anywhere for anything relevant to me that I am aware of. Many references like magazine archives or reference books are held outside the reach of google. I would pay just to be allowed to do the search. Having my local devonthink engine do the search would only need the index to go across the wire which hardly reads very well and so would not signify having made an additional copy of the original. Nor would it likely be problematic to then return snippets or quotes from the text, or metadata. Very obviously to me there could be interesting payment models and features to discuss, but it is far beyond the scope of my suggestion which is that very interesting features could be accomplished with very little work.

* Simple functions for making my own buttons with javascript or whatever would be very interesting as well; add the tag "person" to this newspaper article signifying that I found an interesting person in relation to the topic discussed, perhaps adding the name in relevant metadata if I have it selected, make a pdf, toss it into the selected database/group.

It would need some knowledge about devonthink to follow my train of thought above and I do not have time to spell out a beginners guide here, but certainly ask if anything is unclear or disagreeable. As mentioned there are probably quite a few additional features that could be cheaply and pragmatically had, certainly it would be interesting to hear perspective from other people experienced with devonthink. If any Opera representative would find the topic interesting and would say so, I am quite certain a cross-posting to the devonthink forums would provide as much response and suggestions as one could wish for.