First USA

Join now for free

Search

Products

OSUser
Windows NT
Win95 Reviews
Win95 Tips
Win95 Downloads
Windows 98
MacOS
BeOS
Other
Windows CE
NOSUser

OS News
Site Guide




Advertisement


Also on ZDNet

PC Magazine Labs
PC Week Labs
ZD Internet Labs
Benchmark Tests

Company Finder

Mac
Products
Downloads

At Home

Learning
ZDNet University
LearnItOnline


Magazines

Virtual World

NetMarket




Hacker
BeOS Journal #17:
All the World's a Database

Scot Hacker, ZDNet
02/11/98

If you've had the privilege of witnessing a BeOS demo at a trade show, you know it can be a pretty jaw-dropping experience. It's not unusual for audience members to stand and stare in rapt attention for 30 minutes, suddenly realizing that computing had stopped being fun for them a long time ago... and that it's time to do something about that. People digging into BeOS for the first time often talk about it like they're re-living the thrill of their very first computing experiences. A good BeOS demo can get anyone's adrenaline pumping.

But because the typical trade show atmosphere is so carnivalesque (to put it nicely), Be demos usually feature a lot of whiz-bang simultaneous multimedia stuff, poured on thick to earn the eyeballs of the public quickly. As a result, demo attendees sometimes don't catch some of the deeper, less flashy aspects of the system. They see how powerful and efficient BeOS is, but they don't always see the many ways in which Be is rectifying some of the computer industry's usability shortcomings.

Stuck in the Sand
I'm going to rant for a while about a deep problem with our existing OSs. Bear with me -- we'll get to Be's solution in a minute.

One thing that amazes me about existing operating systems is that they've come so far without figuring out a way to effectively and easily synchronize data generated by different applications. Every time I make a bookmark in Netscape, something in the back of my head wonders if I'm going to need it later while in MSIE, and vice versa. Sure, there are URL synching utilities out there, but they only work if I can find the time to use them. Why is this so? Because Microsoft and Netscape speak different bookmark languages, and don't drop their bookmarks into a shared directory on my system like I've always wanted them to.

It gets worse. If you want to try out a new e-mail program, you've got to figure out a way to bring years of archived mail from your previous program over into a format readable by the new one. Conversion utilities are sometimes available, but they don't solve the synchronization problem -- as soon as you get a single new message, your e-mail clients are out of synch again, not to mention the disk space wasted by duplicate data. And every time I want to try out a new PIM or contact manager, I've got to go through this export/import rigamarole that somehow never seems to go as flawlessly as advertised. And all because these applications refuse, for whatever reason, to share common data formats.

It's weird. We have hundreds of cross-platform, cross-application file formats. Your gif and jpeg, text and rtf, sound and video files can pretty much be hoovered into any suitable application at any time, and you're free to store them anywhere on your hard drive -- not just in the directory of the app that created them. So why in the world aren't there standardized, centralized file formats for e-mail, URLs, and personal/business contacts? Okay, there are a few, like the .VCF business card format, but they're not solving my synch problems, that's for sure. You don't store your images in the same folder with your image editor, so why store your email in the same folder with your e-mail reader? It doesn't make any sense. The vendors have us so locked into their proprietary schemes and data structures that we've become stuck in the sand. After five years with Eudora, trying out another e-mail program is considered "too much of a pain in the neck." Which, of course, is just the way the vendors like it.

The Joy of HotSynching
All of this became painfully clear to me recently when I became a PalmPilot owner. Suddenly -- for the first time -- my home and work computers shared a common database of contact info. Because the Pilot acts both as a portable mini-computer and as a conduit between my work and home machines, I could suddenly see the day's / week's / month's work and home schedules from wherever I happened to be at the time. I could take notes on the train and know those notes would make it into my work and home machines automatically. I stopped emailing myself notes ten times a day. I stopped slapping my head against my forehead every time I realized that some crucial bit of information was "in my other computer." I suddenly realized that I had simply been putting up with non-synchronization because I didn't know there was a solution.

What's the BeOS connection? Be has created a small collection of unique, standardized filetypes for E-mail, People, and URLs. People files in a central directory, ready to be read from or written to by any application. A URL file format, ready to be shared by as many different browsers as you care to install. A common E-mail file format and suggested location, so you can toggle between mail clients at will and never worry about keeping your email apps synched up again. You're free to write scripts or programs that tap into your existing storehouse of email without having to invent a new file format. You can spend two hours deleting unneeded ancient mail from within one program and it's automatically gone from all of them, because there's only one collection of mail files on your system. Which is just the way it should be (this, of course, doesn't mean you're limited to a single email address, as your view onto each address is nothing more than a different Be query -- more on that below).

Data Untrapped
There's no reason for the fun to stop with these three formats, either. You're free to create your own, or suggest new ones to Be (they listen)! Imagine what you could do with a central repository of "Event" files that could trigger system tasks like initiating backups, parsing server logs, or reminding you not to forget Aunt Gertrude's birthday? While on the surface this may not seem so different from existing OS scheduling schemes, the distinction here is that you could treat each Event as an open, independent file that could talk to any other app or to the OS itself, and that can be queried system-wide. The Event is no longer restricted to specialized scheduling software -- by supporting a single, simple filetype, the operating system and every app in it could gain the ability to share a unified virtual calendar -- a common database of past and future occurences to be parsed and accessed in an infinitude of ways. To answer a question like: "Quick -- show me how many network backups have taken place since last Tuesday," you wouldn't have to examine the logs from your backup software -- the operating system would simply "know" by querying file attributes.

How about a Note filetype, which could finally take the notion of the little yellow Post-It deep into the OS, updating the desktop metaphor for the 90s? How about a standard To-Do List file format -- imagine how that could be utilized by various apps for managing personal and team projects, or for posting quick reminders to the screen? These last two are of course inspired directly by the Pilot, which as an example of an incredibly efficient and friendly tool for living, should be viewed as an inspiration to all OS vendors.

There's a big plus here for developers as well as users. They don't have to worry that "people will never give up their legacy of email trapped in Eudora" because it just won't matter anymore. There won't be any reason not to try out a new PIM. Establishing a beach-head with a collection of special filetypes means less risk for vendors and less pain for users.

None of this, by the way, means that developers are constrained to using Be's provided formats -- the OS won't stop you from building proprietary systems. But Be has broken important ground here and provided an opportunity for system-wide openness that most users have never enjoyed. Because all the world's a database and I want to tap it in every way possible, I don't plan to use any BeOS software that doesn't take advantage of these opportunities.


Complete List of
OPERATING SYSTEM REVIEWS

Be Wish List
Operating systems aren't built in a day. Here's a quick list of what's not quite there yet.

Be in the News
Get the latest BeOS headlines from ZDNN
Downloads
PowerMac users, download the BeOS Preview Release 2 for free.
Discuss
Talk with other BeOS users on BeBuzz, the Be message board.
Related
Network OSs

Items below are available at NetBuyer.

Operating Systems

First USA