M's

data search systems

Bell_2

File systems and databases provide different ways of organizing data to help find structure and meaning in what you've stored, but they're not the only approaches possible.  Moreover, the structure they provide is really for one purpose: to simplify accessing it.  Once you realize it's the access, not the structure, that matters, the whole debate changes character. -- Rob Pike Responds

Yesterday I searched for "snort license". Surprisingly, this is what Google returned:

Snort — License: GNU General Public License
According to http://en.wikipedia.org/wiki/Snort_(software)

which means that Google has special code to detect queries about licenses. Nice! (Actually I cringed.)

Data Search

Obviously, the next step is extending this to arbitrary properties, e.g., "rob pike's email", and including synonyms in the search.

There's a joke that there are only two hard things in CS, cache invalidation and naming things (due to Phil Karlton), and resolving such queries is as big a naming problem as it gets. (General question answering is of course harder, but we aren't there yet.)

Basically, you want to be as sloppy as possible in recognizing the query terms, the system should work even if you talk to it on the phone.

For that, we need something even beyond megadatabases, we need data search systems.

Data Search Systems

A data search system answers queries about properties of things (e.g., "elvis' address"). Input, queries, and output are close to natural language.

  • Thanks to the R in REST, data search systems can always fall back to fuzzy results, or add additional information that may or may not be helpful to the user.
  • Data search systems do away with notions about ultra-precise naming and use an acoustic approach to query recognition, and ontological knowledge to find synonyms.
  • Data search systems snarf the whole world wide web, looking for patterns humans use to express data in writing, such as "bedrooms: 2" or "2 bedrooms". (See humane metadata syntax.)
  • Gevil seems to be in the best position to make real data search possible, but technical advances will soon allow a kid with an OLPC to kick their ass.

Dig we must!

April 25, 2007 | Permalink | Comments (2) | TrackBack (0)

plus ca change

By popular demand (hi zanardi!) and crypto-snarkiness (hi chris!), I've reverted the blog back to the old, two-column layout. Anyway, I'm hoping to leave the Typepad silo behind for something better soon.

April 23, 2007 | Permalink | Comments (2)

How rich are desktop apps anyway?

Noguis

Time and time again, I stumble over a blog post that claims that web apps will soon be "as rich as desktop apps". I hope not. The desktop era is over, and good riddance.

Take Twitter. If Twitter were a desktop app it probably would be less useful than it is now. Why?

  • Entering a new item probably would be hidden under File > New and be a modal dialog.
  • I probably would need to save my items regularly.
  • I probably couldn't link to a friend or bookmark items.
  • I probably couldn't open it from my cell phone.
  • I probably couldn't whip up a quick script to fetch or enter some data.
  • I probably couldn't leave or pause the application just by closing a window or opening a new tab.

In these respects, desktop apps SUCK BIG TIME compared to most web apps.

So in what respect are desktop apps richer? Let's look at Adobe Photoshop, a classical rich desktop app. Besides menus and toolbars there's really only one thing that is rich about Photoshop, the canvas. And even that is now available to web apps. Photoshop simply has no visual richness beyond the canvas.

In my opinion, the freedoms (no modal dialogs) and affordances (bookmarks, universal access) offered by most web apps far outweigh any additional richness offered by so-called rich desktop apps. So please, don't drool over some stupid framework like the DOSolate One's Silverlight, which will bring back nothing but a wave of modal dialogs and File menus.

Take the superficial primitivity of HTML, and like Google, del.icio.us, and Twitter, turn it into something extremely useful to millions of people.

April 20, 2007 | Permalink | Comments (4) | TrackBack (0)

The convergence of games and "serious" apps is starting for real... here's a nice network simulator (via chl):

Also:

  • MMRPG meets project management
  • Querying Business Processes
  • "If I drag a coworker to the trash, does that mean he gets fired?"

April 20, 2007 | Permalink | Comments (0) | TrackBack (0)

Core processes require a certain semantic consistency that is certified by a vendor and compliant with authorities. You do not want your General Ledger going out of whack, just because you connected a web service the wrong way to a process that was assembled from multi-vendor pieces. Think of the set of core processes and the master data they require as the part of the brain that takes care of the involuntary actions of our body – like breathing and digesting – you do it, but you don’t think about it, and you definitely don’t want anything messing around with that part of the brain. 

<Shai Agassi>

April 10, 2007 | Permalink | Comments (1)

word

People turn to software to learn the meaning of words, learn which countries were bombed today, and learn to cook a paella. They decide which music to play, which photos to print, and what to do tonight, tomorrow, and Tuesday at 2:00. They keep track of a dozen simultaneous conversations in private correspondence, and maybe hundreds in public arenas. They browse for a book for Mom, a coat for Dad, and a car for Junior. They look for an apartment to live in, and a bed for that apartment, and perhaps a companion for the bed. They ask when the movie is playing, and how to drive to the theater, and where to eat before the movie, and where to get cash before they eat. They ask for numbers, from simple sums to financial projections. They ask about money, from stock quote histories to bank account balances. They ask why their car isn’t working and how to fix it, why their child is sick and how to fix her. They no longer sit on the porch speculating about the weather—they ask software.
(Bret Victor)

April 10, 2007 | Permalink | Comments (0)

My new apartment in Shanghai's busy Jingan Temple area has no internet and I really like it. I finally get to work a bit more. And of course, once you're offline for a couple of days you realize that all those megabytes (gigabytes?) net junkies like me browse every day aren't really necessary or productive.

Two projects I'm still watching closely are Roy Fielding's waka, and Paul Graham's Arc. (Yes, I am also following Project Xanadu.) Despite their vapourware status, waka and Arc interest me, because I love the web and I love Lisp, and both projects promise to make them even better.

Waka aims to replace HTTP's verbosity with binary messages, and replace HTTP's request/response style communication with a richer, asynchronous messaging model (9P anyone?). These two changes should make waka much more useful for networked applications, for example fully Snake Oil, eh, Service Oriented ones, where every component/plugin is a server.

Arc's potential is to revitalize and unify the splintered Lisp scene. I mean, Common Lisp and Scheme environments like SBCL and Chicken are first-grade programming powerhouses which have critical pockets of users, but I think Graham's popularity could trigger something much bigger.

In both cases, it's fascinating that here are two individuals, Roy and Paul, that have the power, thanks to the net, to massively influence the future of a branch of technology, much more than big, faceless corporations could. And heck, maybe even cool stuff like Erlang has a chance to semi-mainstream success, now that there's an excellent Erlang book. You gotta love the net!

March 20, 2007 | Permalink | Comments (3)

one column layouts: the new web designer drug

Now that the early Cyberspace! euphoria is over, people are discovering the joy of ultra-simplicity in web design.

Unlike in print and TV, where fancy layouts have no impact on delivery, complex web designs decrease usability by making the page load slower and feel clumsy.

A recent trend first described by Information Architects Japan is to drop the sidebar altogether and put the other UI elements at the bottom:

Now ask yourself: Where is the user, if he finished the article or if he gets bored. Either he clicks away or he scrolls down to see if there is something coming down there.

Well, in that case, I’d say, provide him with something juicy below the article. It’s the footer you reach when you finish the article. And it’s the footer you scroll to, if you get bored.

Of course, having only a single column is nothing new, but this time, single-column designs are created by people that care about design (i.e. people whose jeans have the current holes/rips/bleaching/coffee stains/whatever.)

Q
WTF?*
Boxblindness
Jakob Nielsen's brain
on three-column layout.
Nyslime
Shows that visual simplicity
and editorial latency
can get you a lot of cred.

See also:

  • "find an old UNIX command that hasn't yet been implemented on the web, and fix that"
  • A widget dork about the atomisation of web content
  • Flickr mobile (mobile use will drive the simplification of web design)

* means What the fuck? in case your grandma asks you...

March 05, 2007 | Permalink | Comments (0)

tradenet.biz, SMS-based African trading platform

ACCRA, Ghana - BusyLab, a software research & development business based in       Accra Ghana, has spent the last two years developing a suite of tools that       will target the agriculture sector across Africa with innovative ways to       market goods and services. The service, known as 'TradeNet' allows users to       sign-up for SMS alerts for whatever commodities and areas they are       interested in. Users can request prices which are provided in realtime on       the network from many market enumerators that are active throughout 380       markets spread across the continent. Users can also indicate their areas of       business and receive instant SMS alerts for offers to buy or sell as soon as     anyone else on the network has submitted an offer via their mobile.

The service online allows analysts to plot trends of historical price data       comparing markets and commodities. Countries can add international commodity       prices, and generate special market reports that can be printed and pasted       in local cybercafes and community markets. Individual users can setup their       own free website with their own custom internet address to advertise their       goods and offers. And farmer and producer groups can setup free websites to       manage all these services and content for their members. Powerful       integration with mobile operators allows anyone to select target users and       publish specific content to thousands of mobiles across the continent. And       within the next twelve months, group loan facilities will be built into the       platform enabling financial institutions to leverage the TradeNet database        and platform to reach the underbanked.

news release, brochure, via del.icio.us/hrheingold

previously: Designing an Architecture for Delivering Mobile Information Services to the Rural Developing World

March 02, 2007 | Permalink | Comments (0)

gasp

Adobe Lightroom makes the baroqueness of current GUIs explicit:

Lightr6_1

This is too much, isn't it?

February 27, 2007 | Permalink | Comments (2)

« | »

Recent Comments

  • Manuel on Helma, the closest thing to a convenient Lisp web app server
  • samantha on Helma, the closest thing to a convenient Lisp web app server
  • Manuel on Helma, the closest thing to a convenient Lisp web app server
  • Christoph Sturm on Helma, the closest thing to a convenient Lisp web app server
  • Kragen Javier Sitaker on innsbruck
  • on eh?
  • Piggywiggs on eh?
  • Manuel on mathematica, micr*formats, and the next unix
  • Tantek Çelik on mathematica, micr*formats, and the next unix
  • Darius Bacon on the little things

Categories

  • Buckybase

About

Subscribe to this blog's feed