Archive for the ‘Ideas!’ Category

All major players in the software industries seem to be interested in the next-generation client/server application. Microsoft used to bet on .NET and wanted everything, from server side to client to be on .NET, except that the most important OS's of the interNET does not have support for .NET; it then focused on the XAML, hoping the new language to break the think wall between traditional client program and web client, when most of the smaller and smarter players have created a lot of wonderful application on the outdated HTML+Javascript technology, based on the XMLHTTPRequest, ironically, introduced by Microsoft. Now Microsoft seems to wake up from its dream of a the next generation. It announced its Ajax tool months ago. Well, that's based on .NET. So it is not fully awaken.

The Web application development tools are really a mess. Different server-side solutions have their own way to simplify ajaxization. Ruby on rail can only be used with Ruby, so PHP people create their own full-stack framework, which is useless for Python fans, who create many yet-another's. We do see people working on JavaScript dedicatedly and generate some interesting work like MochiKit and prototype. But all these tools want you to code JavaScript manually, although the libraries they provide make the programming much easier. After all, the debugging work is still done in this insanely unfriendly language.

Google's GWT really surprised me and made me say "wow". It's really innovative, when Microsoft people are somewhat copying ideas of the JSP tag library. Google's GWT allows your write code in Java and get it debugged on the developer-friendly platform; it treats web browser as a computing platform, and compile your code to that platform. Really neat. We have seen guys implementing a UNIX shell in JavaScript and have never thought about making a subset of Java VM on JavaScript.

The name is interesting, too. The early version of Java comes with AWT, or abstract windowing toolkit; later, the eclipse people builds something called SWT, which is much more powerful and beautiful than AWT. Now Google built this web-based windowing-toolkit-like thing, and the name follows the same trend, although the W is now Web.

Flaws? The GWT approach is very developer-oriented, while the other template-based approach better support the collaboration between developers and designers. With GWT, both the html and the javascript will come from the developer, and the designer have to work from that machine generated HTML code. Must be a bad day for designers. However, when it comes to most enterprise client/server applications, nice-looking is not even a concern. Think about those old-styled VB program with those huge buttons sitting on a grey window frame. GWT's output is much much better than that.


Read Full Post »

I just switched my home page from Google Personalized Homepage to Netvibes yesterday, because of the more content to select from on the later service. I was surprised to see Google released such a wonderful update to its service that I was just switching back my homepage.

It makes so much sense to release an API and to allow customer-made modules to be added to their service. Microsoft did one thing right, that is, to lure developers to their platform. Larger developer population implies more possibilities and choices exposed to the customers, and thus more chances that the customers will choose that platform.

Imagine what a more functional Google Personalized Homepage means. Open the page, all the information you required can be accessed: the weather, the date, your contact book, to-do list, email inbox, selected news, bookmarks to other services, etc. Does it sound familiar?  Isn’t it a decent desktop environment with a dashboard equivalent plus a launcher equivalent? What if there are more online services like on-line word processing, and we create a Mac docker-like module to link to these services? Can we call it web 2.0 OS?

It seems a web platform is emerging, and Google is definitely striving to be the desktop of the new platform.

technorati tags: , ,

Read Full Post »

Web 2.0

Web 2.0 is such a big buzzword now. Having been seeing this word in dozens of slashdot news titles, I have finally made myself search for a definition, after I have seen a list of amazing Web 2.0 applications.

Unfortunately, the defintion in wikipedia is still too vague and not technical enough. Thanks to the Web 2.0 API website, I finally have some basic understanding about this excitingly promising creature.

The most important feature of Web 2.0 , I believe, is the capability to resemble UNIX pipes. Based on XML, web sites, or web services, can communicate and integrate, making web-based application capable to cross the site boundary and be much much more functional.

A necessary and important complement to this essential feature is the capability to present the content to user with a rich user interface, which is created with Ajax and/or XUL and/or Flex in the current Web 2.0. I really hope XUL can take the world, because I think Ajax is, fundamentally, a hack. A hack should be a temporal workaround rather than a long-term solution.

Allowing users to contribute back to web is also essential, coz that is why the Web 2.0 should contain much more interesting stuff and thus be more fun than its predecessor.

technorati tags: , , , ,

Read Full Post »

I guess I should use emacs to do wiki, but VIm is so great at keeping me in its reign. For some daily but simple note taking, I have been using Zim. But up till now, I have found several unpleasant retrictions.

  1. No tag, no search! The only choice to use the namespace mechanism to orgainze. Such an approach is not as flexible (or intuitive) as tag-based system;
  2. No more than 1 format applied to every text region. You choose to underline it, then forget about use italic text or monospace font. That’s a big problem;
  3. Export is too bad!
  4. No customization of font, background, etc. It should at least use CSS to allow customizable view.

technorati tags: ,

Read Full Post »

Ideas about flock bookmarks

Flock has a nice bookmark management system: favorites. Instead of using an hierarchical directory structure for storing and organizing the bookmarks, flock hires a tag-based approach. I think this approach, coz now you do not have to go through a deep tree and/or pick from a long long list. Just search it.

However, I am uncomfortable about 2 things in this bookmark system:

  1. Bookmarks will always to sent to the onling bookmark system, if you turn that feature on. You are not given the fine control of whether a bookmark should be synced and shared in public. The coarse control only allows you to use or not use that feature, applied to all bookmarks.
  2. No shortcut. I like the shortcut-for-bookmark feature in Firefox and Mozilla. It allows me to type in the name of a bookmark to get the fastest access. I use “news” to go to my favorite new site and use “w stuff” to search the “stuff” in Wikipedia. It’s unlimintedly expandable for me. It’s easy to embed this feature into the Flock favorite system (just by storing the shortcut name in the description). Really hope they can make it happen.

technorati tags: , ,

Read Full Post »