Friday Nov 12, 2004

Microsoft seeks to out-search Google

Microsoft puts its dukes up to the practically omniscient Google with it's new search engine out to beta. Competition is good for us web seekers, but if you are a heavy search engine user, you can use searchenginewatch to keep a balanced view of the internet. Alternatively, you could try the charmingly named Dogpile meta-engine to look for some new search engines. But my favourite "meta" method for finding search engines to date has to be the Google Labs 'Sets' feature with search criteria like  "google" and "lycos"!

What can we expect next from search engines? I think the 'Sets' approach has huge potential (especially if it's backed by a semantic network with weighted nodes, optionally cross-referenced with significant terms from the Glossary), but I'm also betting on: grammar correction, federated travel search, dynamic directory (get a yahoo style tree, but based on a semantic analysis of the search terms), RSS meta-search, ... but that's enough for one blogity.
Steve Ballmer looks ready to go for Google

Wednesday Nov 10, 2004

Firefox goes prime time

Great to see the Firefox browser get to the magic 1.0 today, and nice to see that users are beginning to abandon IE to get something that is clean, fast and best of all secure (the spyware fiasco that Internet Explorer wrought among users was amazing; meanwhile Microsoft were offering partner anti-spyware companies a listing on their IE security pages - somehow I don't think that's the right way to create a market opportunity! ;)

I use Firefox across all the desktop operating systems I run, and I'm looking forward to seeing Firefox bundled with Sun products soon.

From a developer standpoint, the best part about Firefox is that it's lightweight enough to bundle even with standalone desktop products; this will really help the Gecko/XUL engine to get deployed, and the fact that Gecko is a common base between the Mozilla and Firefox browsers makes it an easy platform to test against.

Now all we need is a UNO/XPCOM bridge to unite the middleware in with and we have a very compelling open-source developer platform spanning Windows, Linux and Solaris - nice. UNO components are already programmable in C++, Java and Python (not to mention all the Java-hosted languages out there; Groovy, Jython, BeanShell and so on). And of course with JDIC and the SDK you can use Mozilla or as "mega-beans" within a Swing application - awesome.

Friday Oct 15, 2004

UI design - improving on Cooper's personas

I've just read George Olsen's article Making Personas More Powerful: Details to Drive Strategic and Tactical Design; superb content and new ideas that seem grounded in broad experience.

Also on the Boxes and Arrows site, there are plenty of HCI-related articles, including more practical ideas on personas, usability studies, and user's mental models.

Not only is the quality high and the noise level low, but the articles have a standard style and layout that makes them easy on the eye - it's well worth setting some time aside to browse the other topics too.

Tuesday Oct 12, 2004

StarOffice (or as a Blog-editor?

I'd like to ask your opinions on a matter close to all our hearts ;)

There are some who think that StarOffice or should be a Blog-editor (it could also be a Wiki-editor, or a better HTML editor, e.g. with better support for css styles).

What do folks think about making StarOffice a Blog-editor? Is this

a) a core feature of StarOffice, OR
b) should a separate Blog-editor application be able to use StarOffice as a rich-text editing component, OR
c) a combination of both, where a document may be either edited in a separate app or posted to a blog using a menu nearby File/Save or Export?

I lean towards b) (for example, 2 years ago we did a prototype of StarOffice used as an editor in the Java IceMail mail client), but I would welcome your comments!

Saturday Oct 09, 2004

Feeling vulnerable?

SysAdmin's and security buffs are already reading this (warning: it's a biggie but a goodie) which is an annually posted summary by a global team of experts that identifies and solves the top vulnerabilities impacting computer systems. And hopefully the new chief of cyber security in the US will get a mandate to tackle the criminals who have been exploiting them (btw, why was that a BBC link and not one of the many other news articles that cover this? simple - no ads, and there's even a low-graphics version - nice one).

Wednesday Sep 15, 2004

Lego -> Social and "Fictional" Software

When I read Paul Lamere's blog 27 Stages of Lego Sorting, it reminded me of Douglas Coupland's wonderfully insightful book Microserfs which is about a group of developers (some of whom had major Lego phases, one of whom goes on to create a Lego masterpiece ;) who form a kind of geek commune around a company to develop a software application called "Oop!". The idea of "Oop!" was to provide a rich kind of virtual Lego set, with the ability to use, build and share smart mobile & reactive components (a bit like Lego Mindstorms), which could act like doors or lifts or game characters or Rubik's cubes or vehicles, and so on. Ok, pretty interesting idea.

But that started me thinking; that's not the first time I've seen a software application described in a novel. Douglas Adams in Dirk Gently's Holistic Detective Agency also outlines a rather amazing piece of software which is best described by Douglas (2) himself:

'Well, Gordon assigned me to write a major piece of software for the Apple Macintosh. Financial spreadsheet, accounting, that sort of thing, powerful, easy to use, lots of graphics. I asked him exactly what he wanted in it, and he just said, "Everything. I want the top piece of all-singing, all-dancing business software for that machine." And being of a slightly whimsical turn of mind I took him literally.

'You see, a pattern of numbers can represent anything you like, can be used to map any surface, or modulate any dynamic process -- and so on. And any set of company accounts are, in the end, just a pattern of numbers. So I sat down and wrote a program that'll take those numbers and do what you like with them. If you just want a bar graph it'll do them as a bar graph, if you want them as a pie chart or scatter graph it'll do them as a pie chart or scatter graph. If you want dancing girls jumping out of the piechart in order to distract attention from the figures the pie chart actually represents, then the program will do that as well. Or you can turn your figures into, for instance, a flock of seagulls, and the formation they fly in and the way in which the wings of each gull beat will be determined by the performance of each division of your company. Great for producing animated corporate logos that actually mean something. 'But the silliest feature of all was that if you wanted your company accounts represented as a piece of music, it could do that as well. Well, I thought it was silly. The corporate world went bananas over it.'

Reg regarded him solemnly from over a piece of carrot poised delicately on his fork in front of him, but did not interrupt.

'You see, any aspect of a piece of music can be expressed as a sequence or pattern of numbers,' enthused Richard. 'Numbers can express the pitch of notes, the length of notes, patterns of pitches and lengths.' 'You mean tunes,' said Reg. The carrot had not moved yet. Richard grinned. 'Tunes would be a very good word for it. I must remember that.'

(aside: I think Douglas Adams had several important insights into software, one of which is quoted at the end of an interesting article Autistic Social Software (not really about autism, but about the ADD of software products and/or developers who don't want to try to understand the social implications of networked communication and collaboration; if you're into groupware though, my single favourite groupware article is one that I read 10 years ago - What do groups need? A proposed set of generic groupware requirements - it offers an excellent synthesis and framework for evaluating groupware requirements; gosh, back then I used e-mail to request a copy from one of the authors Munir Mandviwalla; he kindly sent me a hardcopy which I still re-read occasionally - great stuff, and now you can get it on the ACM portal; both of these articles only go to show that creating truly new forms of groupware is very hard)

I bet that's just the tip of the iceberg though in terms of creative software ideas in fiction; anyone know any others?

Tuesday Sep 07, 2004

Bridges to/from RSS

That reminds me; how well are developers served by technologies that bridge to RSS (that is, they read (or write) RSS and write (or read) another format/protocol)? Well, there's:

  • nntp//rss (RSS to NNTP; Java)
  • IM/RSS bridges, including JabRSS (RSS to XMPP; Python)
  • mail/RSS - including rss2email (RSS to SMTP; Python)
  • rssutilities (RSS to HTML; Java) - a JSP tag library for adding RSS feeds to a web-page
  • Rome (RSS to/from other versions of RSS and Atom; Java) - RSSLibJ seems to be the nearest alternative in terms of features, but Rome nudges into first place in my book due to it's depth of RSS support and extensibility - Alejandro Abdelnur and  Patrick Chanezon are co-authors (alongside Elaine Chien) with blogs

So compared to the number of RSS client applications (such as the rather nice RSSOwl) or add-ins, there are actually not too many real bridge technologies, but at least API's like Rome make it easier to write them.

Monday Sep 06, 2004

More thoughts on Bookmarks

I noticed RichB's thoughtful Thoughts on Bookmarks blog - you don't need to wait for Google labs to solve this, just use Google search "collaborative bookmark" and you can see the wealth of work that's already been done in this space! I agree with Rich that one valuable feature of shared bookmarks is the fact that they are maintained by a group, but I think a collaborative bookmarking system can have some other equally useful attributes:

  • bookmarks can be checked regularly for update or availability by a network server
  • each bookmark may be linked at multiple nodes in a taxonomy, or even in multiple discrete taxonomies
  • you could subscribe to be notified to changes in specific taxonomies or categories
  • if the structure of the taxonomic hierarchy is a separate entity, it can easily be shared which has two benefits:
    • users don't have to learn multiple taxonomic structures and the labels for folders in the hierarchy
    • a user could create a federated view of multiple taxonomies by a subscription mechanism; this would allow you to subscribe to a shared bookmark taxonomy with folks who have similar interests or a compatible way of classifying bookmarks
  • could be capable of associating a title, description and keywords with a URL (the network effect of sharing a URL encourages the user to provide more detail)

A collaborative bookmark manager would complement the linking style of blogging nicely; blogs are great for providing topical links that have a short half-life, whereas a shared bookmarking tool would help to create more long-term value.

I think such a tool could have a natural fit with content management systems (CMS); instead of persisting transient web content to your own hard disk, you could request that a web-addressable page or document would be cached in your group's CMS (subject to legal limitations on storage due to content licensing). If you access a bookmark via the bookmark server, it could automatically redirect to the local cache.

Naturally there are issues of importing and synchronising bookmarks from existing end-user tools such as Mozilla or IE, but none of that is rocket science. It would also be desirable to be able to sync with public bookmark taxonomies like the DMOZ Open Directory; that could be helped if DMOZ exposed an RSS interface to complement it's HTML one.

Of course, RSS is a pretty effective way to expose the structure, content or updates of any collaborative bookmark taxonomy.

Follow-up: RichB comments that he was really thinking about a personal bookmark manager, but given his stated issues (1. the time it takes each person to setup this set of bookmarks that can help them with their day to day work. 2. knowing what that magic set of best bookmarks are.), I don't see an ideal solution for an individual. However if an individual's bookmark taxonomy had some structure (categories or folders) in common with a shared bookmark source, it would be possible to merge selected categories/folders of the shared bookmarks, either dynamically or a one-time static merge.




« July 2016