“[…] This database stuff is clearly the next frontier. If we are ever going to have natural speech interfaces and virtual assistants that collect and display our information in more useful ways, the computer is going to need to know much more about how our files relate to each other and remember what we’ve done to them and who’s done it. Moving all this organization to the system level instead of locking it up in separate applications that are then duplicated in a search index seems like the best way to get there.”
How Much Better Can OSX Get?
About The Author
Eugenia Loli
Ex-programmer, ex-editor in chief at OSNews.com, now a visual artist/filmmaker.
Follow me on Twitter @EugeniaLoli
So far, whenever software attempts to anticipate my next move, it ends up as nothing more than giving me a headache. I don’t believe this can change any time soon. Being able to access the recent files I used with a certain application from the File menu is the best it can do – it’s consistent, predictable and non-intrusive. Anything other than that I either can’t imagine or it’s not feasible.
Take word – the least recently used menu items “disappearing” is annoying. It’s unpredictable and ends up costing me more time when I need a menu item that was there a month ago, but is now hidden. It would be far more convenient if it just let me edit those menus, instead of trying to do that for me.
So far, whenever software attempts to anticipate my next move, it ends up as nothing more than giving me a headache. I don’t believe this can change any time soon.
Agreed. I’ve grown to admire the Classic MacOS posthumously – it sure had its technical flaws, but there was a minimalist ease-of-use philosophy behind it that Douglas Adams summed up as “There is no problem so complex that a simple solution cannot be found to it.”
Whereas the approach taken by Windows (and, increasingly, OS X) is to build abstractions over an a complex underlying system in order to shield users from those underlying complexities (paradoxically, adding more abstractions exacerbates the problem by making the OS even more complex). Or as Adams put it, “You have a complex problem, so here is a complex solution.”
Take word – the least recently used menu items “disappearing” is annoying. It’s unpredictable and ends up costing me more time when I need a menu item that was there a month ago, but is now hidden.
Ugh, agreed again. It’s annoying enough for proficient users – and whoever conceived of that particular “feature” has clearly never spent any time training novice computer users. It boggles the mind that ill-conceived features like that can make it into Windows, yet Explorer still lacks a “resize window to fit contents” command or a proper keyboard shortcut for making a new folder.
What is with these articles from people who just speculate about crap and no background worth two s**** to draw from?
Is there some special high school click I should know about?
Clique.
I agree with tyrione. The article is just some guy’s speculation about what would be ubercoolwtfbbq. Nothing really groundbreaking. Nothing to add, no envelopes pushed, no revolutionary ideas. You should add an osnews category entitled, inspired by tyrione’s post, “highschool fanboy blog posting,” so we know which articles not to read.
Personally I travel alot with my laptop, I am looking fwd to the day which GPS devices will be embedded into laptops for meta tagging information so that I can ask the computer to pull up all the files I worked on in DC or something. Would be quite usefull… and for the love of god, they need partial word matching in spotlight.
And I wonder (now that it seems most all the new intel macs have built in isights) how you could work a project like this:
http://www.no-more-blind-computer.com/
into the primary os? use your facial expressions and/or actions in front of the computer as ways to tag files. Like show me what I was working on when I was eating breakfast on tuesday? … I know I am dreaming, but it’ll happen someday I am sure.
BTW, I am not holding my breath for real voice recognition until another 5-7 years out.
That would actually be cool.
Unlike certain steps BACKWARD like vocal interfaces. Don’t get me wrong, voice recognition has its uses, and interfacing with your PC everyday is a bad one.
For your PDA though…
I can type faster than I can talk, as can many people, and I can edit a thousand time faster than I can say how to!
“I can type faster than I can talk, as can many people”
I am sure most people can talk faster than they can type.
I am sure most people can talk faster than they can type.
Well for a non-native (english), I can type much faster and better than I can speak. For my native language it might be different though.
Speaking of vocal interfaces, why is voice recognition seemingly going nowhere? Controlling my Media PC with a remote can be slow at times – I want a microphone mounted near the TV that picks up my instructions.
I dont want to wear a headset, I want arbitrary speech recognition (so I can say, say the name of a recorded television show I want to watch), and preferably no training ahead of time.
Sure I dont ask for much
But it seems there’s been almost no progress in this area for desktop PC’s.
What’s most likely to happen (already happening) is that mobile phones will be GPS enabled. And since they support bluetooth you should easily be able to connect your computer with them.
Other than that a phone-sized bluetooth GPS receiver now is less than 100$ on ebay an d you can use it with today’s pcs. The most annyoing thing with GPS is that you have to be outside with a clear view to the sky for it to work. Well, it could record your last position before going dark.
Federal law in the US requires all phones to be GPS enabled and broadcast it when the phone is in emergency mode (911 dialed and no phone restart)
Its been in affect for 5 years or so, so everyone who gets new phones every 5 years or so should have GPS in their phones by now.
I am looking fwd to the day which GPS devices will be embedded into laptops for meta tagging information so that I can ask the computer to pull up all the files I worked on in DC or something. Would be quite usefull… and for the love of god, they need partial word matching in spotlight.
Me, I’m looking forward to the day when my computer stops hanging. Having a fancy new way to access my files might save me time, but it can’t save more time than I lose every day waiting for applications to respond, and sometimes relaunching them after a crash.
It’s silly to spend programmers on new features when they could be fixing problems. Especially considering how much anger is produced by problems, and none by lack of new features; nobody hits their computer and curses because it is missing an integrated persistent object store.
Caj
pfft, I do that every day.
While software prediction (up until now) has been entirely infuriating, I would love it if my computer was smart enough to do something like this, for instance …
Let’s say I’m going through a list of old, non-present devices in Windows’ Device Manager, doing a right-click, Uninstall, OK. After 5 or so times that I do this, the computer should be smart enough to notice (after watching me) that I’m removing non-present devices from my system. It should then pop up a dialog, with something along the lines of “It looks like you’re doing ______. Would you like me to do the rest for you?”
Sure thing, Mr. Computer!
But no.
That’s more about the inefficiencies of batch processing in a point and click gui. This is where the power of a command line, and a good scripting environment can be invaluable. Rather than the computer guessing what you want it to do (which is guaranteed to be hopeless), you tell it exactly what you want it to do.
“It should then pop up a dialog […]”
“It looks like you’re writing a letter, would you like me to:
1. Help by presenting a series of useless “wizards”
2. DIE! DIE CLIPPY, DIE!!!!!!”
—
[2]
Talking paper clip: “It looks like you are writing a letter!” I think I’ll pass. I’d rather have something like this:
A view of all devices known to the system. A column that shows the last date/time the devices were used. Click on the column header to sort the list. Select multiple devices. Click a button to remove all selected devices.
!!!
Well, some of the ideas are good, since the ease of use, nice applications with good integration, and petty tricks in the appplications is what make OS X an ergonomic, useful, and good OS.
So, improving that with additions to the OS or the applications is always good.
Other things presented in the blog are mere speculations of a Mac fan.
But IMHO, if you want to know “How Much Better Can OSX Get?”, I would say that still pretty much…
Network protocols, and network and server security and speed have to be improved.
Also server applications, and a general improvemnet in speed would be highly appreciated.
Better recognition of other file systems, and more interoperatibility with other Operating Systems would improve OS X and would attract more users.
These are specific areas in which OS X, which is currecntly a nice, solid, and good OS could and should really improve…
!!!
Edited 2006-04-04 05:11
Just like they always do.
Apple and Microsoft both don’t invent anything.
They just borrow and then redesign.
Kind of like George Lucas and the StarWars franchise.
http://www.jitterbug.com/origins/lightsabers.html
I remember an old cartoon joke that had a dinosaur sleeping and read:
“Millions of years ago, dinosaurs invented rubber tires. Since cars didn’t exist yet, they used them as pillows.”
🙂
There is something to be said about combining several existing technologies and making a “whole” that properly leverages them.
Metacrap: Putting the torch to seven straw-men of the meta-utopia ( http://www.well.com/~doctorow/metacrap.htm ) is a nice rant about the problems with metadata.
People always seem to forget collecting everything but the simplest of metadata is hard work. As someone who tagged and organised a 30Gb mp3 collection I can vouch for that. Even with automated tools it is still a pain. Just uttering the magical words “database” isn’t going to solve anything.
Edit: my favorite quote from the above rant “Short of breaking fingers or sending out squads of vengeful info-ninjas to add metadata to the average user’s files, we’re never gonna get there.”
Edited 2006-04-04 09:38
Which is the point – Most people can’t even be bothered to organize their files into folders and give said files meaningful names, and people like the author of the article expect them to fill out all this other crap too? One look at a directory full of S2010001.jpg S2010002.jpg etc, etc, off a digital camera is proof enough of that. There’s a reason people like browsing as thumbnails.
The only reason it even comes CLOSE to working with MP3’s is the automated tools and name repositories – and then that only works because those are commercial files (or illegal copies therin) that would be the same for all users. The minute you start trying to apply this nonsense to every little file the user creates be it a word document, spreadsheet, picture off a digital camera, etc… It’s just not going to fly given the laziness of the average user.
Seriously, who the {censored} do they expect to fill out all those extra fields when most users can’t even be bothered to use filenames bigger than 16 characters long?
You make a good point. However, one could get a long way with simple metadata, if it’s automated.
For example, I wish I had ‘Downloaded from:’, ‘Downloaded when:’, besides ‘Last modified’ data, on files I download via Firefox or Emule (with Emule the files could just indicate that it’s downloaded via Emule).
Another thing I wish I had an attribute-capable OS for: checksums as attributes. That’s the sort of thing that could easily be implemented via a filebrowser plugin or a simple standalone app, IF the OS I use had advanced attribute capabilities.
“Short of breaking fingers or sending out squads of vengeful info-ninjas to add metadata to the average user’s files, we’re never gonna get there.” [awesome quote] might be true, but we can get at least halfway there.
“You make a good point. However, one could get a long way with simple metadata, if it’s automated.
For example, I wish I had ‘Downloaded from:’, ‘Downloaded when:’, besides ‘Last modified’ data, on files I download via Firefox or Emule (with Emule the files could just indicate that it’s downloaded via Emule). ”
yeah a
from: GAIM: brendan
from: seamonkey: http://www.archlinux.org/download.php
from: pacman: community (or whatever package manager, be nice for spot cleaning a filesystem)
from: diskdrive: vactionpics (not always helpful but should be there fore completeness)
would be really handy
For example, I wish I had ‘Downloaded from:’, ‘Downloaded when:’, besides ‘Last modified’ data, on files I download via Firefox or Emule (with Emule the files could just indicate that it’s downloaded via Emule).
I don’t know. Maybe it’s my paranoid streak but I’m wary of my computer automatically collecting data on my actions. You never know what information can be abused by a boss or angry significant other, let alone by police (especially for those in more undemocratic countries)
Another thing I wish I had an attribute-capable OS for: checksums as attributes. That’s the sort of thing that could easily be implemented via a filebrowser plugin or a simple standalone app, IF the OS I use had advanced attribute capabilities.
I guess on Windows you could put all that info in an alternate data stream (on NTFS). The problem would be getting all aplications to use it.
“””For example, I wish I had ‘Downloaded from:’, ‘Downloaded when:’, besides ‘Last modified’ data, on files I download via Firefox or Emule (with Emule the files could just indicate that it’s downloaded via Emule).
I don’t know. Maybe it’s my paranoid streak but I’m wary of my computer automatically collecting data on my actions. You never know what information can be abused by a boss or angry significant other, let alone by police (especially for those in more undemocratic countries)”””
if you have porn on your work computer your boss will probably be more concerned with the porn than the site it came from, and ideally any metadata would leave the system along with the, it really isn’t more revealing than your browsers history, and again ideally you’d be able to mass edit metadata manual so if you say have a bunch of stuff off emule you could go to search type:
from: emule
CTRL+a
rightclick edit metadata, from: ‘computername’
and it would be indistinguishable from something you made yourself
Uhh… did you try iTunes?
I know my piddly 18.44GB music collection is nothing compared to your 30GB, but it was no hassle at all to find the albums/tracks that didn’t have proper tags and update them. Just select a whole album -> get info, put in the name and album title. Then you do just click on the title and F2 if it didnt guess it properly.
The best part? Easy to maintain. all I have to do now is to make sure that new music that gets merged into the library has proper tags.
Don’t like iTunes? Okay, whatever, use MusicBrainz or something to do it for you.
If the OS supplied the MetaData for you, and all you had to do was add your notes to it? It would rule. “Metatopia” might not exist, but it is possible to use metadata effectively. Just ask my ipod.
Uhh… did you try iTunes?
Well I started before iTunes, before I had a mac actually, but you’re right iTunes did make it easier to update some tags. I can also recommend MP3bookhelper for windows users ( http://mp3bookhelper.sourceforge.net/ ) It’s open source too.
Don’t like iTunes? Okay, whatever, use MusicBrainz or something to do it for you.
Yeah but I have obscure punk cds, local band demo’s and Belgian music that isn’t in any of the internet metadatabases.
Dude, I was out last night and I saw OSX at the bar. That guy is awesome! We had a few drinks and I ran out of money and he totally bought all my drinks for the rest of the night. Then Windows came in and called us a bunch of losers, what ev! Later Windows drank so much he passed out in the corner and someone put his hand in a glass of warm water. Ha ha ha! Serves him right. I was too drunk to drive so OSX even gave me a ride home and it was totally out of his way. Anyways, OSX really is so much better than Windows!
I would like to see OSX run on any Intel box. That would make it a lot better instead of buying proprietary Apple overpriced hardware.
I must admit that my eyes glazed over very quickly when trying to read the linked to blog, so I only skimmed it…
Apple had a fully integrated database-based storage system 10 years ago. They also had an extensible intelligent assistant system that would make use of metadata stored within the databases. You could give the assistant a command such as “schedule lunch with Ford Prefect at Milliways tomorrow”, and it would put an appointment in your diary linking in card records for both Ford Prefect and Milliways. And if you happened to have more than one Ford Prefect or Milliways in your address book it would ask you which one you meant. That’s just one example of the kind of smart stuff that was possible, and as I said this system was extensible. You could potentially ask the assistant “give me directions to Milliways” and (if you had a GPS and navigation system installed) it could sort that out for you.
It was all very smart stuff, and worked exceptionally well.
Only problem was that it was on the Newton OS, which Steve J in his infinite wisdom decided to kill off.
Unfortunately no effort has been made to date to integrate this kind of technology into Mac OS. The only part of Newton that has survived is the handwriting recognition engine.
I’ve often thought that it would be great to get the intelligent assistant back and the close integration of core apps that Newton had and put into Mac OS X. However the big problem with this is that applications have to be (re-)written to use the universal database storage system.
First the valid point that corrupting an iPhoto database can lose all the metadata in it… then a call for implementing the database at the system level, where corrupting it not only loses the data but may wreck the entire OS.
This is what happened to Windows with the Registry: granted it’s not Windows’ only problem, but the registry concept (which probably sounded good to an executive) stands as one of the worst design decisions ever made.
My chief objection to WinFS is the fear that something could go wrong and I’d be left with a crashed system and tens of thousands of GUID-named files in a single huge directory. I give Apple slightly more credit than that, but it’s not unlimited.
Not that a database file system with reasonable performance wouldn’t be nice to have when it was working. Most of what he wants (metadata, less the standardization and universal search/access) was available with MacOS resources years ago, and it was great.
How Much Better Can OSX Get? – If only it supports the Apple ‘multi-touch’ Graphics User Interface:
http://video.google.com/videoplay?docid=6379146923853181774
(The Multi-Touch Interaction Research website: http://mrl.nyu.edu/~jhan/ftirtouch/ )
some nice ideas about search, tough they actually just tell us what a conceptual linking engine would do, like KDE is goin’ to have for KDE 4 (tenor).
http://www.osnews.com/story.php?news_id=10295
…that given enough time, Apple will eventually re-create BeOS.
Moving all this organization to the system level instead of locking it up in separate applications that are then duplicated in a search index seems like the best way to get there.