From what I can tell, Spotlight and Google Desktop are very similar. Both are system services that incrementally index changes in binary files, using custom parsers for each supported format. Spotlight has an advantage in that it's an extensible architecture and supports more formats out of the box. Spotlight could conceivably also have better index freshness, considering Apple controls the entire software stack. But for the most part, you're talking about a fairly easy-to-implement service, assuming you have access to decent parsers for common data formats and a good full text search engine (like Apache's Lucene).
WinFS, at least conceptually, is a totally different animal. Rather than files being opaque, linear sequences of bytes, files are composed of schematized data, much like XML documents and their XSD schemas. So much like XML documents in large part obviate the need for application developers to create their own binary formats (and the necessary binary parsers and writers), Windows developers can write to these higher-level APIs.
Furthermore, WinFS allows the operating system to have much greater access to the semantics of the data, and as such, it can provide more services to the developer. For example, WinFS schema allows you to express DAG relationships between entities, doing ref counting and cleanup when an entity is no longer referenced. Also, you can get event notifications not just at the file level, but on sub-file-level entities. The WinFS team was even going to provide sub-file-level replication "for free", but I believe this may have been shelved for now.
And they have also put a fair deal of work into letting all this work alongside the "legacy" datastream model.
From a full-text-search perspective, probably the biggest difference vs. Spotlight and Google Desktop is that any native WinFS-based app should not need to write a plugin. The relevant plaintext and metadata would simply be transparent to the filesystem.
Of course, at the moment WinFS is far from shipping while Google Desktop is out and Spotlight seems imminent. Still, WinFS is as ambitious as anything I've heard of Microsoft doing, and as far as I know they're not following in anyone's footsteps this time.
Here is a decent high-level summary of some of the features of WinFS. You can also go to MSDN and check out the preliminary API's (I believe they're all under the System.Storage namespace).
Don't get me wrong, Apple does cool stuff but their strongsuit is marketing, not "invention".
You've got to give them credit for product design as well. Nobody makes more desirable-looking software and hardware. Is it any wonder that Apple's fiercest supporters are graphic designers?
There's nothing radical about this, technically speaking. Do a search for IFilter, which is Microsoft Indexing Service's way of doing the same thing. (And there's no way MS was the first to do this, either.)
Unfortunately, Microsoft chose not to expose this functionality to Windows users. Which is odd, considering IFilters for Office documents are installed on every Windows (2000+) machine, as is Indexing Service.
Microsoft provides this for.NET, in the form of the "ngen" tool. Although it isn't commonly used by developers (nor does MS really encourage them to), when you install the.NET Framework, the standard libraries are "ngen-ed".
(FWIW, my company tried ngen to speed up our app's startup time, and it made surprisingly little impact.)
Outlook Express lets you drag messages out to the filesystem; the raw RFC822 (headers, attachments, and all) gets saved as a file with an.eml extension. When you double-click the.eml file, OE displays it in a message window.
Thunderbird lets you save by pressing Ctrl+S on a particular message, but it doesn't work when you have more than one message selected, and you can't use drag-and-drop either. And I haven't been able to figure out how to display the message in TB again (there's a File | Open Saved Message..., but it doesn't seem to do anything).
See, you would think e-mail would be a world of nice, standardized protocols and perfectly compliant mailers and transport servers. The sad reality is that SMTP, RFC822 headers, and MIME are abused or incorrectly implemented all over the place, often in more subtle and insidious ways than HTML. (To be fair, some of the standards are pretty hard to write a 100% compliant parser for.) This is not too surprising when you think about how many e-mails probably get generated by sloppily coded CGI scripts and spam engines.
Thunderbird does a pretty good job with the standards, from what I've seen, but it's not perfect (unless they've made changes for 0.9). For example, it doesn't always provide the correct representation of e-mail addresses, specifically those that use quotes in the name. From what I remember, it also does not do as good a job of parsing illegal MIME as Outlook Express does.
But those are very small niggles, interesting only to someone who is implementing his own e-mail/MIME parser (for example). In real-world usage, it would be pretty unlikely to run into legitimate messages that TBird couldn't parse.
Re:What about performance and memory usage?
on
Thunderbird 0.9 Released
·
· Score: 4, Interesting
I don't think it's necessarily fair to compare Thunderbird to a web browser. TB has a lot more data to juggle than Firefox, in general. FF just has to deal with a couple of webpages at a time, while TB has to keep giant lists of messages at the ready. Also, just in terms of raw amounts of data, your average mbox has a lot more data than your average HTML page.
I've found TB to compare favorably, performance-wise, to other clients I've tried, such as Outlook, Outlook Express, Evolution. (Although it's been a long time since I've tried Evolution.)
I don't know about the American Civil War, but we only entered WWII after Japan attacked us at Pearl Harbor. Historian Arthur Schlesinger, Jr. calls the issue of whether or not the US should enter WWII (before Pearl Harbor) the most divisive national issue he has seen in his lifetime--and he's old. Not even the Vietnam War could match the ferocity of the debate around WWII.
In fact, Schlesinger's recent book says that the US has had strong isolationist tendencies for most of its (short) history--for example, our failure to join the League of Nations, when the sitting US president was the driving force behind it. It's only been since the end of WWII that America has embraced... uh... what's the opposite of isolationism? Multinationalism?
Not to say that the US should return to an isolationist approach to foreign policy. Just pointing out that it's far from a given that we have a moral obligation to cross the globe and usurp dictators who do not pose an immediate threat to our national security, no matter how badly they treat their own people.
And usually good always wins in the end, and the world is better off.
Does good usually win, or is it just that the winners get to write the history books?;)
"And I tell you, God only knows, that we never had the intentions to destroy the towers."
He meant before America sided with Israel against Lebanon and Palestine. He says so, and then continues, "And as I was looking at those towers that were destroyed in Lebanon, it occurred to me that we have to punish the transgressor with the same -- and that we had to destroy the towers in America so that they taste what we tasted, and they stop killing our women and children."
What he's trying to say is "Yeah I took out the WTC, but you started it."
error in sql-statement: mysql_connect, err-no: 2002 description: Can't connect to local MySQL server through socket '/var/lib/mysql/mysql.sock' (111)
Warning: mysql_select_db(): Can't connect to local MySQL server through socket '/var/lib/mysql/mysql.sock' (111) in/srv/www/htdocs/web0/html/epiacenter/includes/sql_ layer.php on line 89
Warning: mysql_select_db(): A link to the server could not be established in/srv/www/htdocs/web0/html/epiacenter/includes/sql_ layer.php on line 89
error in sql-statement: mysql_select_db, err-no: 2002 description: Can't connect to local MySQL server through socket '/var/lib/mysql/mysql.sock' (111)
error in sql-statement: mysql_query, err-no: 2002 qry: select active, view from nuke_modules where title='Content' description: Can't connect to local MySQL server through socket '/var/lib/mysql/mysql.sock' (111)
Warning: mysql_fetch_row(): supplied argument is not a valid MySQL result resource in/srv/www/htdocs/web0/html/epiacenter/includes/sql_ layer.php on line 298...
But that potential will remain unrealized unless Firefox gets some kind of additional client-side programming platform. Again, one of the open source JVMs, but without Sun's bloated class libraries, could do the trick. JavaScript by itself really just doesn't.
Just curious if you've actually tried XUL + JavaScript. I've done some Swing work, a ton of C#/Winforms work, and about a month and counting of XUL/JS work, and so far the XUL/JS experience has actually been pretty good. You might be surprised how much you can get done in the XUL without dropping into JS at all; in Winforms and (especially) Swing, the general purpose programming language (C# or Java) is responsible for declaring the UI, which most Swing developers will tell you is just incredibly painful. (Winforms isn't as bad because of RAD tools, otherwise it would be.)
So for many applications, the amount of JavaScript you have to write is pretty small compared to the amount of C#/Java. And anyway, JavaScript isn't a half bad language anymore; C# and Java could learn a couple of tricks from it (closures in particular are invaluable for GUI programming, and neither C# nor Java have them yet).
I thought this too, based on plenty of experience writing DHTML in JavaScript 1.1/1.2, but I was recently forced to write a Firefox extension (using JavaScript 1.5) and it was surprisingly painless.
Unless one of your criteria for a nice language for big and complex projects is static type checking (or static anything checking, I guess), you might be surprised how far JavaScript has come along. It actually has a coherent OO model (though it's different than the class/instance model used by most mainstream languages these days), it's got closures, exception handling... many programmers have written complex things in languages that gave them much less.
We get a http://web.mit.edu/loginname/ directory, and the already extant directory/loginname/www/ is by default world-readable; but they don't set up Web pages for us. (You can also request a static IP address & hostname and run your own server. Yay for self-sufficiency!) In the strictest sense, though, I guess you could say they don't provide an easy interface for setting up our page (just ftp & kerberized telnet.)
When I got to MIT in 1996, you had to use MH to check your mail. It didn't matter if your major was computer science or music--if you wanted to check your e-mail, you had to to inc, scan, show, repl, rmm, and comp from a bash shell. (This was a most humbling experience for me, as a college freshman who had worked as a part-time IT helpdesk technician since I was 12 years old.)
The day MIT offers an "easy interface" for students to set up a personal web page, I will shed a nostalgic tear.
Slashdot Mirror
WinFS, at least conceptually, is a totally different animal. Rather than files being opaque, linear sequences of bytes, files are composed of schematized data, much like XML documents and their XSD schemas. So much like XML documents in large part obviate the need for application developers to create their own binary formats (and the necessary binary parsers and writers), Windows developers can write to these higher-level APIs.
Furthermore, WinFS allows the operating system to have much greater access to the semantics of the data, and as such, it can provide more services to the developer. For example, WinFS schema allows you to express DAG relationships between entities, doing ref counting and cleanup when an entity is no longer referenced. Also, you can get event notifications not just at the file level, but on sub-file-level entities. The WinFS team was even going to provide sub-file-level replication "for free", but I believe this may have been shelved for now.
And they have also put a fair deal of work into letting all this work alongside the "legacy" datastream model.
From a full-text-search perspective, probably the biggest difference vs. Spotlight and Google Desktop is that any native WinFS-based app should not need to write a plugin. The relevant plaintext and metadata would simply be transparent to the filesystem.
Of course, at the moment WinFS is far from shipping while Google Desktop is out and Spotlight seems imminent. Still, WinFS is as ambitious as anything I've heard of Microsoft doing, and as far as I know they're not following in anyone's footsteps this time.
Here is a decent high-level summary of some of the features of WinFS. You can also go to MSDN and check out the preliminary API's (I believe they're all under the System.Storage namespace).
Try Google Desktop Search, and see how completely the resource consumption problem has been solved in 2004.
You've got to give them credit for product design as well. Nobody makes more desirable-looking software and hardware. Is it any wonder that Apple's fiercest supporters are graphic designers?
Unfortunately, Microsoft chose not to expose this functionality to Windows users. Which is odd, considering IFilters for Office documents are installed on every Windows (2000+) machine, as is Indexing Service.
(FWIW, my company tried ngen to speed up our app's startup time, and it made surprisingly little impact.)
What in the world does revenue have to do with anything? Shouldn't it be about selection, price, codec/bitrate, and possibly DRM technology?
Thunderbird lets you save by pressing Ctrl+S on a particular message, but it doesn't work when you have more than one message selected, and you can't use drag-and-drop either. And I haven't been able to figure out how to display the message in TB again (there's a File | Open Saved Message..., but it doesn't seem to do anything).
See, you would think e-mail would be a world of nice, standardized protocols and perfectly compliant mailers and transport servers. The sad reality is that SMTP, RFC822 headers, and MIME are abused or incorrectly implemented all over the place, often in more subtle and insidious ways than HTML. (To be fair, some of the standards are pretty hard to write a 100% compliant parser for.) This is not too surprising when you think about how many e-mails probably get generated by sloppily coded CGI scripts and spam engines.
Thunderbird does a pretty good job with the standards, from what I've seen, but it's not perfect (unless they've made changes for 0.9). For example, it doesn't always provide the correct representation of e-mail addresses, specifically those that use quotes in the name. From what I remember, it also does not do as good a job of parsing illegal MIME as Outlook Express does.
But those are very small niggles, interesting only to someone who is implementing his own e-mail/MIME parser (for example). In real-world usage, it would be pretty unlikely to run into legitimate messages that TBird couldn't parse.
I've found TB to compare favorably, performance-wise, to other clients I've tried, such as Outlook, Outlook Express, Evolution. (Although it's been a long time since I've tried Evolution.)
Microsoft Outlook 2003 offers them as well.
Looks like it's--let's say "inspired by"--the Bang & Olufsen BeoSound 1, which is itself no paragon of hifi value.
(true story)
Wow, do you really think the opposite of isolationism is imperialism?
The word I was trying to think of was internationalism.
In fact, Schlesinger's recent book says that the US has had strong isolationist tendencies for most of its (short) history--for example, our failure to join the League of Nations, when the sitting US president was the driving force behind it. It's only been since the end of WWII that America has embraced... uh... what's the opposite of isolationism? Multinationalism?
Not to say that the US should return to an isolationist approach to foreign policy. Just pointing out that it's far from a given that we have a moral obligation to cross the globe and usurp dictators who do not pose an immediate threat to our national security, no matter how badly they treat their own people.
And usually good always wins in the end, and the world is better off.
Does good usually win, or is it just that the winners get to write the history books? ;)
He meant before America sided with Israel against Lebanon and Palestine. He says so, and then continues, "And as I was looking at those towers that were destroyed in Lebanon, it occurred to me that we have to punish the transgressor with the same -- and that we had to destroy the towers in America so that they taste what we tasted, and they stop killing our women and children."
What he's trying to say is "Yeah I took out the WTC, but you started it."
Actually I believe "Favorite Equation..." is in the lead. See?
Astronomist and cartoonist... a true Renaissance man.
explanation in plain english
Yeah, here's the text:
/srv/www/htdocs/web0/html/epiacenter/includes/sql_ layer.php on line 89
/srv/www/htdocs/web0/html/epiacenter/includes/sql_ layer.php on line 89
/srv/www/htdocs/web0/html/epiacenter/includes/sql_ layer.php on line 298 ...
----
error in sql-statement: mysql_connect, err-no: 2002
description: Can't connect to local MySQL server through socket '/var/lib/mysql/mysql.sock' (111)
Warning: mysql_select_db(): Can't connect to local MySQL server through socket '/var/lib/mysql/mysql.sock' (111) in
Warning: mysql_select_db(): A link to the server could not be established in
error in sql-statement: mysql_select_db, err-no: 2002
description: Can't connect to local MySQL server through socket '/var/lib/mysql/mysql.sock' (111)
error in sql-statement: mysql_query, err-no: 2002
qry: select active, view from nuke_modules where title='Content'
description: Can't connect to local MySQL server through socket '/var/lib/mysql/mysql.sock' (111)
Warning: mysql_fetch_row(): supplied argument is not a valid MySQL result resource in
Just curious if you've actually tried XUL + JavaScript. I've done some Swing work, a ton of C#/Winforms work, and about a month and counting of XUL/JS work, and so far the XUL/JS experience has actually been pretty good. You might be surprised how much you can get done in the XUL without dropping into JS at all; in Winforms and (especially) Swing, the general purpose programming language (C# or Java) is responsible for declaring the UI, which most Swing developers will tell you is just incredibly painful. (Winforms isn't as bad because of RAD tools, otherwise it would be.)
So for many applications, the amount of JavaScript you have to write is pretty small compared to the amount of C#/Java. And anyway, JavaScript isn't a half bad language anymore; C# and Java could learn a couple of tricks from it (closures in particular are invaluable for GUI programming, and neither C# nor Java have them yet).
Unless one of your criteria for a nice language for big and complex projects is static type checking (or static anything checking, I guess), you might be surprised how far JavaScript has come along. It actually has a coherent OO model (though it's different than the class/instance model used by most mainstream languages these days), it's got closures, exception handling... many programmers have written complex things in languages that gave them much less.
When I got to MIT in 1996, you had to use MH to check your mail. It didn't matter if your major was computer science or music--if you wanted to check your e-mail, you had to to inc, scan, show, repl, rmm, and comp from a bash shell. (This was a most humbling experience for me, as a college freshman who had worked as a part-time IT helpdesk technician since I was 12 years old.)
The day MIT offers an "easy interface" for students to set up a personal web page, I will shed a nostalgic tear.
Oh, here's the MIT page. I think with those two "No" answers corrected, MIT should be #3 on the list.
Does the school provide Web pages?
Can students register online?
Both these answers were "No" according to the survey, but they should be "Yes".
It's a little ironic that you wrapped "Just Use Python! :)" in curly braces, isn't it?