Slashdot Mirror
Microsoft's Decade-old Patent On Tree-view Mode!
BhaKi writes "Remember the Tree-View mode in many file management applications? It's shocking to know that this omnipresent feature was patented by Microsoft back in 1995 (granted in 1997). I'm not very sure about the implications, though. The patent is so general that it can be related to many things from tree-mode to virtual filesystems. Check out claim no. 3 of the patent for the most clear part."
I patented trees in general in 1992. I'm going to sue.
You can get patent results ad-free from Google or straight from the USPTO.
http://www.google.com/patents?id=uRkgAAAAEBAJ&dq=5689662
Hell, its vague enough to cover slashdot's hierarchical nesting message view. Pay up, slashy!
Table-ized A.I.
It appears we have stunning commentary on the nature of software patents. If a company, "first posts," if you will, they get the right to license the patent to others and sue those that don't pay them royalties. Truly, parent is a modern genius of metaphor.
Thank god for mc and ztree
provided a navigable file system browser for DOS - tree view of directory in the left pane, list of files in the currently selected directory on the right.
It was a best-selling product, too.
A few years later (but before 1995), IIRC Lotus Notes had a navigational pane in its client. I doubt it was even the first app to use that - it was just sort of common wisdom among UI designers at the time.
I have to give MS kudos for not using this patent offensively.
There is a war going on for your mind.
There's so much prior art here it's not funny. For example, Executive Systems first published XTree for DOS (later XTreeGold) in April, 1985. It was the absence of this functionality in MS/DOS that make the functionality so popular. This is just another example of how the software patent system is truly messed up and needs (and hopefully will get) a serious overhaul.
Generally, very little. Yes, most low-level things in CS have been patented in some sense (XOR cursors, one-click checkout, run-length image encoding, multi-hash lookup, stacktrace error display strategies.)
In theory, all software development grinds to a halt. In practice, no one gives a damn.
Trying to enforce a very broad software patent usually just gets the entire patent invalidated. Even if you win, you get to play whack-a-mole with a thousand open-source projects. And most software is bespoke stuff within corporations: good luck tracking that down to enforce patent claims.
Unless you are a law firm with the business model of extorting cash for infringment, you lose by going to court. Bad press, skeptical judge (unless you are suing a direct competitor,) workarounds from the peanut gallery provided pro-bono, countersuits from others with overlapping clainms: it gets ugly fast. Better to just cross-license and get on with life.
That's interesting, because a virtually identical view was available in a Wang Laboratories software product called Clearview, released in 1989, which ran on Windows 2.0.
(Clearview was one of a genre of Windows add-ons, HP NewWave being probably the best known, that plastered improved graphics shells or desktop managers on top of Windows).
And Clearview itself was nothing more than an improved version of a directory display that was used in the Wang Laboratories OIS circa 1977. They were logically the same, although visually different because the OIS was constrained by having a character-oriented screen. At least within Wang itself, Clearview's directory display was regarded a spiffy bitmapped graphic version of the OIS's display.
I seriously doubt that Wang was first or even close to first, but Wang was definitely shipping large numbers of commercial products that offered tree views of directories long before 1995.
"How to Do Nothing," kids activities, back in print!
Surely a family tree would be considered prior art?
Then why can't I use a mouse to drag my mother-in-law to the recycle bin?
Table-ized A.I.
This patent isn't just about trees, or even file-system trees (which Microsoft made prior art for with the old Win3.1 File Manager). This is about file-system trees that also include things that aren't actually in the file system. It's about how things like Control Panel and My Network Places can appear in the same Windows Explorer tree with your C: drive.
Hopefully, though, the whole thing is now moot.
(T>t && O(n)--) == sqrt(666)
Perhaps Slashdot story submitters should have to certify that their understanding of patent law comes from something more than perusing the musing of RMS at the League for Programming Freedom. A short course in claim construction would help in differentiating between broad patents that cover something commonly used and narrow patents that are easy to avoid.
The claims of this patent all include the limitation of a "name space extension" that adds at least one "non-file system object" to the file system display. For example, opening a zip file as a folder is an example of adding a non-file system object. If it was part of the shell, instead of an extension to the shell, then the shell probably would not infringe. The capability of extension by third-party vendors is one of the distinguishing characteristics of Windows Explorer.
There may be prior art that invalidates this patent (after all, prior art can be in any language, any time before invention). But, it would have to be prior art that reads on the narrow scope of this patent, not the bizarre, broad interpretation offered by the submitter and by /. commentators.
This is about file-system trees that also include things that aren't actually in the file system.
You mean things like devfs, procfs, and tmpfs?
I notice a lot of people tend to make really really really generic patents these days, presumably to give them more control over who they get to sue, but when it comes to Prior art, does just one instance invalidate the whole patent, or does it only invalidate certain aspects of it?
+1 IDisagreeSoHeMustBeATrollOrAnAstroturferOrAShill
prior art from 1986 Plan 9 V 1.0
There are places where the networks are not touching,and there are places where they are-Boeing's Lori Gunter
as a node at a level where it is not in the FS
You know, there is this old family of operating systems dating back from the early 70s, that tend to represent pretty much everything as a file system, even things that aren't necesarily on the disk like processes or more recently USB devices.
"Sufficiently advanced satire is indistinguishable from reality." - [Tips: 1DrYakQDKCQ6y52z6QbnkxHXAocMZJE61o ]
The little known computer language Mumps (Massachusetts General Hospital Utility Multiprogramming System - http://www.cs.uni.edu/~okane) implemented all these claims 30 years earlier. Documentation from that earlier era pretty much shows all the features that M$ appears to be claiming. Another case of M$'s deja vu all over again?
Kevin O'Kane http://www.cs.uni.edu/~okane/
Burroughs (now Unisys) CANDE (command and edit language) had this feature in 1969.
Thanks for clarifying that. Still OS/2 2.0, released in April 1992, had the ability to display "non file system object". Any (former) OS/2 developer (or user with knowledge of low lever working) will tell you that what you just described it is the main feature which makes the OS/2 workplace shell the best GUI shell ever.
As for handling ZIP: IBM never bothered but it was available to the WPS as third-party extension.
Martin
I don't see the relevance of "well UNIX has been placing all devices in the filesystem for decades".
FYI, in addition of "/dev/" which *is* a directory on the file system (but contains special files to communicates with the hardware), there are other things that *are not* in the file system.
The patent covers including things which aren't in the filesystem as part of the tree view.
There are entry point in the tree like "/proc" which aren't in the file system at all.
Instead they use a special module called PROCFS (it's a file system drive in unix world, and pretty much equivalent to the the extension that the patent mentions : both can be used to make structures not on a disk appear as directory-like tree)
and expose a complete directory-like structure which in fact doesn't represent files, but represent processes, kernel status, etc.
See the above post for a reference on a paper about how exposing processes with an interface that look likes files.
In short : if there's a directory called "/dev/snd/" this directory exists because it is on the EXT2 partition mounted there. The directory "/proc/kernel/" exist because the procfs modules makes it available. It has no existance on the disk.
Similar mapping of abstract concept into something that "looks like a directory tree" is very popular in unix.
More recent example are pts which is used to represent the various (virtual) terminals in directory-like fashion structure. usb exposing the topology of usb devices. Also sysfs in linux is used to represent pretty much anything internel of the kernel like system drivers.
Most of the example I give are recent (usbfs and sysfs are post 95), but they are representative of a tendency that has existed in Unix for a long time because it was in its design.
The patent covers the software mecanism which covers the possibility that was introduced in Windows 1995:
Using a simple file browsing software like Explorer, you access a single tree structure which can hold both actual elements on the disk like "Desktop -> My Computer -> C:" (gives access to a physical partition), and things that are actually abstract element made visible in the same tree "Desktop -> My Computer -> Control Pannel -> {some settings}" (whose functionality is coincidentally is pretty much close to what Linux's recent "sysfs" or parts of the mid-80s old unix "procfs" where created to do).
(And somewhere in between, the case of network resources, which are remote resource made visible in the local tree : Windows' "Desktop -> Network Neighbourhood -> {variable number of indirections depending on Windows version} -> {server} -> {ressource}" is exactly functionally equivalent to a unix' mounted remote file system)
So in short, my opinion about the patent :
"Congratulation, you've successfully described something that has been in Unix for the past 25 years"
----
In fact Unix' original implementation is much closer to what the patent describes than Windows 95 (that microsoft where trying to patent) : /proc really exposes things that where never part of the file system to begin with.
Speaking of system settings and internal data
Whereas the functionally equivalent "Desktop -> My Computer -> Control Pannel" trick simply lists all ".cpl" files in "\\windows\\system[32]\\". It's not an extension that make out-of-filesystem object visible in the tree, it's simply a thing that acts as a filter for files already elsewhere on the same system.
In addition, the old unix implementation I am referring to is better integrated in the OS. Anything made available in the tree using a special filesystem driver that exposes abstract things instead of actual filesystems, is instantly made available to any of the usual tools (including command-line tools) used to ha
"Sufficiently advanced satire is indistinguishable from reality." - [Tips: 1DrYakQDKCQ6y52z6QbnkxHXAocMZJE61o ]