Instiki - Devonthink Pro

Consider this message a succesor to the posting … php?t=1882 .

I remember quite well how excited i’d been after having read the following paragraph within the Technology section of the DEVON-Technology webpage:


Some of the most likely applications for the DEVONtechnology are databases (knowledge bases, expert systems, search engines, table-of-content-generators, instant data-mining), intelligent agents, encryption, compression and archiving. Other uses for DEVONtechnology may be chat-bots, context-sensitive help, optical character recognition (OCR), speech analysis, recognition and synthesis, statistics, thesauri and automatic translations.[/i]

… and my curiousity grew into impatience with version 1.7 of DEVONThink. The company and its prodduct started to exceed its promisses. At the same time it was clear that some kind of scripting or programming support was needed to turn the information retrieval tool into an artificial intelligence development environment as outlined on the website.

Now the Pro version obviously can be considered as a next step into the soft computing future. But Applescript, despite its merits has the big disadvantage not to be a perfect fit for a world of a web-operating-systems. For exactely this reason i’ve built my Ruby-DEVON-Bridge, which uses the power of a multi OS compatible scripting language with an extraordinary soft computing tool.

Two flies with the clap of one hand !

Suddenly i could connect Devonthink to all the cool webtools popping up everywhere around the web and secondly everything performed much better than expectd.

One of the first apps developed with the combo is an extended version of Instiki, the well known wiki-clone, written in Ruby on Rails, under Mac OS X. My intention was to develop a workflow which should make it possible for me to screen a much bigger amount of RSS, E-Mail and IM information flow than usuallay capable for an individual. Basically the main idea is to establish some kind of smart twist. Instead of searching and screening data for hours make the appropriate information find you, automatically. This is possibel with Devonthink.

So, what i did was to build a workflow on top of several applications, namely Gmail, Jabber, NetNewsWire, Instiki and Devonthink. Via RSS or Atom i collect all my news-feeds, blog-feeds, tag-feeds. photo-feeds, mails and im-sessions into NetNewsWire. Every our a script colleccts the as unmarksd flagged messages into Devonthink. Everything into one big data-soup. i don’t even try to screen this.

Next i’m using my favourite wiki, Instiki. There i simply enter everything that belongs into notebook. The daily ToDos, collect ideas, copy-paste my articles, collect links and pdfs, post addresses and events. Then something very interesting and effective happens. The same moment i post a new wiki page it’s entry gets compared to the whole datasoup inside Devonthink. So should there be a relevant RSS-Feed concerning my actual thoughts or intentions i will see it and can call it directly. That’s a total eye opener. Things become relevant when i articulate myself or when i log or describe relevant working practices. Documents start to group around my postings automatically. The high performance desktop becomes a reality, this way. This way 500 rss feeds are screened every day, the whole wikipedia, all my mail for 3 years and much more.

To make the similarity results even better i intend to filter the replies in accordance to the folder hierarchy inside Devonthink, by using a simple rule-based engine to decide the relevant context of every posting.

Finally, here’s a screenshot visualizing the whole thing:

There is another one showing how simple it is to push things into Devonthink with Ruby:

And another one showing how raw Apple Events are used to embed Devon-intelligence into Ruby ( not that simple anymore but performance is fantastic) :

And dont’t forget ! Instiki means this tools can be used by a group of users, if you use instiki as a wiki on a public webserver !

More to come…

This sounds really exciting though I don’t fully understand what you’ve done from your brief explanation and the screenshots. I look forward to hearing more about where this is going to go for the rest of us who aren’t capable of writing our own custom software, and note that Eric sounded favorable in principle in the thread you refer us to.

Indeed! 8)

johnsilver, it took me a few readings to follow what you’ve done. I’m not a developer, though I have surface familiarity with Ruby/Ruby on Rails. Is this right: you’re using Ruby on Rails for inter-application communication between Instiki and DTPro. Specifically, when you make an entry in Instiki, DT’s “see also” function is invoked, and the results are piped back to the Instiki page, automatically creating a relevant set of links related to the original posting. (I told you that I’m not a developer :smiley: )

This is some seriously cool functionality – providing DT’s advanced functions as service to other applications. Please keep us posted on your progress.