Unlike electric motors, computers do not have so easily quantified characteristics that might correspond to a motors "size" and "power".
Computers are defined not only by their computing capacity, but by their I/O devices, and the availability and organization of nearline storage. As networking, interface and storage technologies change, so will the computing devices.
Because its ultimately whats going in and out of the computer(s) that matters.
Many people will still own general purpose PCs in the future, if only to take advantage of the latest trends in storage and A/V/brain-interfacing technologies. These technologies will adapt for the most general purpose devices first, before expanding to specialized personal information devices and console-like entities.
I'm not saying they're isn't going to be segmentation/stratification in the market based on user sophistication or needs, but you're still going to see "general purpose" personal computers for a very long time.
...do us a favor. Don't ever link Gibson Research on Slashdot ever again. Steve is such an attention-grabbing tool. And you want us to sit around and listen to a fucking podcast? Jesus christ, do you think we have nothing better to do? A quick HowTo or Wiki is just fine, thank you.
But yeah, MAC address filtering has no purpose other than to frustrate you when you use a new network adapter.
I didn't mean to sound confusing. The NT4 PDCs are used for small ad-hoc networks not part of the corporate LAN (lab facilities). The nice thing about Samba is that you can do things like force NTLMv2 even if NT4 didn't support it, for extra security. Its great when we can have a single-signon type password for Windows, Linux and Solaris in a lab environment.
What's really wicked is that you can also do a Windows 2000 PDC-like thing with LDAP and Kerberos if you're feeling real adventurous, but the extra effort probably isn't worth it since they don't got much in the way of Group Policy editors for the benefit of the Windows boxes. It's an interesting excercise if you want to really "scale up" or muck around with lots of extra LDAP schemas to do other things... but uh the NT4 emulation is good enough for a less than 100 computer, 100 user setup.
I've got all my Unix boxes at work joined to our AD 2003 forest. And I've got Unix boxes emulating hybrid NT4 DC/Sun NIS domains My Evolution talks to the Exchange server... and our intranet is firefox friendly.
We have a lot of Dell server hardware and it doesn't seem to fail any more or less often than any of our other equipment. That being said we _never_ buy PERC anymore... we've been burned by it too many times. Software RAID over 3rd party SCSI controllers is plenty fast for most jobs... if its not we go buy RAID offload enclosures from Xyratec or somebody.
PERC... good hardware plus shitty Dell firmware. What were they thinking?
Wave Race fucking owned. That game was so much fun to play... and all my friends try to race me and they're all like WTF HAX because all they do is sit around and play GTA all day. lmao
I must concur that Diet Mountain Dew has the most pleasant and consistent flavor of any diet drink (when compared to its sweet counterpart). Actually, I prefer the Diet variety to its full-sugar cousin, the sweetness is not as intense and it has a more pleasant citrus flavor. Very little aftertaste, essentially none. Alas very few places other than supermarkets and convience stores carry it. The only place that I know of that has it in fountain form is Quiznos (which is a good excuse to go there). Even more hard to find (and even better tasting) is Diet Mountain Dew Code Red. They don't even stock the 20oz in 7-eleven anymore.:-(
This is true. It doesn't necessarily make it less crash prone. But it does make it instrumentable if it proves to be unstable (you could easily trace, debug, intercept, or otherwise validate the requests the blob made if so needed).
Furthermore, the kernel mode portion would merely be relaying commands to trusted memory-mapped regions and IO space requested by the process initially (limited by configuration files, perhaps). Most kernel crashes are the cause of errors (pointer mistakes, buf overflow, race condition, etc.) in the complex driver code which "trap" the system in kernel space. The user space portion would likely instead SIG11 and die... if it left the hardware in a weird state it could be fixed by simply restarting the driver program which would, at its outset, send RESET type commands to the device putting it in a known state.
The largest problem I see is that it isn't possible to easily recast a userspace driver program into a device node without a mechanism like FUSE. It only works if the hardware target in question is nearly always accessed behind a userspace library (OpenGL, libalsa/libjack/OpenAL, libusb).
Most drivers don't need to run in kernel mode (read: any USB device driver)... or at least they don't need to run in response to system calls. The hardware manipulating parts kernel should stick to providing higher-level APIs for most bus and system protocols and provide async-io for kernel and user space. If most kernel mode drivers that power your typical/dev/dsp and/dev/input/mouse and such could be rewritten as kernel-threads that dispatch requests to and from other kernel threads servicing physical hardware in the system you can provide fault-isolation and state reconstruction in the face of crashes without incurring much overhead. Plus user processes could also drive these interfaces directly so user space programs could talk to hardware without needing to load in dangerous, untrusted kernel modules (esp. from closed-source hardware vendors).
Or am I just crazy?
Yeah but microkernels seems like taking things to an extreme that can be accomplished with other means.
Dell and HP don't sell "propietary" service manuals to authorized service representatives. Any document the person on the forum who would have cared to link to a supposedly incriminating page could have linked directly to Dell's or HP's support sites.
Nearly _every_ computer and computer component is manufactured/assembled in China. The exceptions are the large-chip manufacturing fabs run by big chip giants in Germany, Malaysia, India, and other semi-technologically savvy areas with cheap labor.
Rip your subtitles to a vobsubout.sub/.idx pair. If you check the.idx file, you can see what colors it intends to use on playback in the header file. You can change them there (they're just like HTML colors, #RRGGBB). Your.idx may be broken if you only get white text on playback. Check to make sure it's not trying to display a subtitle image at a weird time (often happens when ripping across cells when selecting chapter ranges)... that can confuse the playback software and cause it to miss settings or not display subtitles at the right times. I often have to hand edit the.idx files and tweak them myself.
Then again it could be an artifact of encoding the subtitles into the resulting avi. I don't use that method anymore, ever.
I encourage you to not encode the subtitles in the movie itself... but bundle it along (either as seperate files or as part of an OGM). It allows the end user to change color/transparency, move it, display it elsewhere (in the black bars going from widescreen->full, for example), and it improves encoding performance.
transcode (and then also mjpegtools) are kinda more of the attitude: do exactly this one thing and no more... for use in big conversion scripts and the like.
mencoder is more like: I just want you to convert this into this with this restriction and I don't care how you do it.
Of course, sometimes it guesses wrong about what you want it to do. Getting it do things a certain way can require black magic and animal sacrifice.
BSPlayer does not link or bundle in a full ffdshow library. It can leverage the ffdshow DirectShow filter to play a lot of media types without using other WM/DS libraries (people often prefer the features of ffdshow in MPEG2/MPEG4 over filters bundled with DVD drives and/or DivX). Usually you find BSPlayer and FFDShow bundled together, for example, in the KLite Codec Pack.
However, BSPlayer is a much better parser of video container formats (ASF, WMV, AVI, OGM) and MPEG transport streams than most other players out there (maybe with the exception of VLC). All of them are better than any versions Windows Media Player.:-/
So it can handle broken, badly indexed, or partially downloaded files with ease. Additionally, like VLC, mplayer and MPC, it can handle extended features in video containers that many other players (Windows Media Player included) omit. For example, multiple video streams, subtitles, multiple audio tracks, etc.
General Purpose complex CPU is slower than Single Purpose parallel CPU at the operations the Single Purpose CPU is designed for!!! Film at 11
I mean are you fucking kidding me? Banias, Dothan, and Sledgehammer do the GPP thing just fine at the same power/performance level as PPC or USPARCIV.
Of course NVidia would love to be put out their own CPU. Because then they could control THE WHOLE SOLUTION! CPU, chipset, video and sound! All they'd need is a HD manufacturer to partner with...
It's not the x86 dragging them down, its having to make their chips compatiable with the products of the rest of the industry. Big friggin deal... welcome to the modern tech marketplace.
No modern webbrowser allows cookies to "leave a trail" of where you've been anymore. Cookies are only sent in requests to the domains that created them in the first place.
I guess one valid point would be that ad-networks can sort of track you by determining what sites who use their banners you visit based on referrals headers. The issue there is: 1) There's no user-identifiable information that links your cookie to the pages you visited feature that adnetwork's banners aside from your IP address at the time 2) Adnetworks don't care _who_ you are. They use the cookie to determine what sites should be considered "related" so they can better determine how to distribute banners. (That is, banners that do well on one site might do well on sites visited by people who visited the first site) 3) You can always block referers to images/iframes hosted on external domains (adnetwork stuff). IIRC this is the default now on Firefox and Safari... so... its irrelevant now.
* EPIA small/home theatre CPUs and mini-ITX boards * AMD and Intel chipsets (K8Tx00, PT8x0, CX700, etc.) * basically any southbridge not made by NVidia or Intel * Envy24 audio chips (lots of cards, from Hercules to Terratec to M-Audio) * VIA Velocity GBE and Rhine 10/100E ethernet chips (used in many, many desktop cards) * A metric fuckton of small desktop switch integrated platforms... and a bevy of wireless MACs
Plus the S3 video chipsets are used in a whole lot of video-on-board motherboards even if the north/southbridges AREN'T from VIA. Basically if your server doesn't have an ATI Rage-XL or Intel IGP on it, then its got an S3. Tons of HP and IBM rackmount systems have them.
Unix systems allocate a fixed area in each "cylinder group" to store the inode tables. Which means it doesn't get fragmented. OTH you can run out of inodes (or, conversely, you can give it too many and waste potential file storage area). NT has a dynamic allocation scheme but since the MFT is a file too it can get fragmented if you let your drive fill up.
That being said, probably the only file that matters performance-wise if its fragmented is the MFT on a Windows System. Fragmentation is not a big a deal as it seems... what's more of an issue is the availability of coalesced free space in which to write new data.
Unix systems don't have defragmentation tools because the allocation schemes are usually simpler and degrade nicely, and also because fragmentation for reads is pretty much irrelevant since on a multiuser system the drive head is going back and forth all over the disc anyway. Typically the system queues up all the pending IO requests and reorders and issues them in an elevator up-and-down pattern to maximize throughput.
If your unix system gets so fragmented that single-user performance is suffering and writing files takes to long, the "answer" is to tar up a subdirectory, or use find with a size>Xkb restriction, and copy them somewhere else, delete it, and then put it back on the volume. You can even do that to system files while the system is up since unix doesn't care when files get used/replace that are currently being used (unlike Windows). That is just as good as an in-place defragment.
1) Remote desktop == Remote assistance which means that you can help people with their spyware problems remotely without walking them through a Timbuktu or GoToMyPC install (blegh)
2) Volume Shadow Service == Restore Points, driver rollbacks, and backups of your system drive that actually work.
3) Those locking improvements are really important now that everyone has hyperthreading or dual core machines. It makes applications quite a bit snappier interactively, and helps when your doing stuff like burning to a CD while streaming a movie from another drive.
4) The 64-bit memory map stuff is very important because it removes a lot of stupid limitations that you ran into before. When you started seeing systems with > 1GB of memory the system wouldn't page it as effectively and the system was swapping a lot more. Also if you work with multimedia/content creation apps there were limitations to the file size you could use for memory mapped access (which is more useful when doing editing when you both seek and stream a lot... so developers would have it fall back to the shittier open/read/write-style access API).
And did I mention a built-in bluetooth and wireless stack?
These changes had a lot more implications in 5.1 over 5.0 than you give them credit for. I find myself frustrated when going from a 2003/XP system to a 2000 system when things I expect to be there or for it to be able to do aren't there or are done in a less sane way.
once you strip it down. I don't bother and just install server 2003 (same thing really)
XP adds over 2000: *) Volume Shadow Service (finally...) *) MFT defragmentation support *) 64-bit virtual memory support and 64-bit file mmaps *) application prefetch *) better page retiring algorithms and PTE management *) new (faster) locks for kernel objects for SMP *) Built-in Terminal Server and detachable console (nee remote desktop) *) support for read-only NTFS (like on a CD or ZIP disk) *) redesigned registry and configuration subsys (much faster, less mem, mostly lockfree) *) slow int 0x2e syscalls replaced by sysenter/sysexit (big difference in context switches)
and a bunch of other little things behind the scenes.
So its worth going to NT 5.1/2 if you can whip it into shape.
If you go to yours or another users's page, you can see how they tagged articles (and whether they replied). Presumably this is a quick way to get a feel for what a user thinks about articles. I understand the ability to search for articles tagged a certain way is also forthcoming.
Slashdot Mirror
Unlike electric motors, computers do not have so easily quantified characteristics that might correspond to a motors "size" and "power".
Computers are defined not only by their computing capacity, but by their I/O devices, and the availability and organization of nearline storage. As networking, interface and storage technologies change, so will the computing devices.
Because its ultimately whats going in and out of the computer(s) that matters.
Many people will still own general purpose PCs in the future, if only to take advantage of the latest trends in storage and A/V/brain-interfacing technologies. These technologies will adapt for the most general purpose devices first, before expanding to specialized personal information devices and console-like entities.
I'm not saying they're isn't going to be segmentation/stratification in the market based on user sophistication or needs, but you're still going to see "general purpose" personal computers for a very long time.
...do us a favor. Don't ever link Gibson Research on Slashdot ever again. Steve is such an attention-grabbing tool.
And you want us to sit around and listen to a fucking podcast? Jesus christ, do you think we have nothing better to do? A quick HowTo or Wiki is just fine, thank you.
But yeah, MAC address filtering has no purpose other than to frustrate you when you use a new network adapter.
I didn't mean to sound confusing.
The NT4 PDCs are used for small ad-hoc networks not part of the corporate LAN (lab facilities).
The nice thing about Samba is that you can do things like force NTLMv2 even if NT4 didn't support it, for extra security.
Its great when we can have a single-signon type password for Windows, Linux and Solaris in a lab environment.
What's really wicked is that you can also do a Windows 2000 PDC-like thing with LDAP and Kerberos if you're feeling real adventurous, but the extra effort probably isn't worth it since they don't got much in the way of Group Policy editors for the benefit of the Windows boxes. It's an interesting excercise if you want to really "scale up" or muck around with lots of extra LDAP schemas to do other things... but uh the NT4 emulation is good enough for a less than 100 computer, 100 user setup.
cp /tmp/install.$$/my_script > /etc/rc.d/init.d/my_script_name /etc/rc.d/init.d/my_script_name /etc/rc.d/init.d/my_script_name /etc/rc3.d/S00my_script
chmod a+x
ln -s
I've got all my Unix boxes at work joined to our AD 2003 forest.
And I've got Unix boxes emulating hybrid NT4 DC/Sun NIS domains
My Evolution talks to the Exchange server... and our intranet is firefox friendly.
So uh... why do I need Microsoft again?
We have a lot of Dell server hardware and it doesn't seem to fail any more or less often than any of our other equipment. That being said we _never_ buy PERC anymore... we've been burned by it too many times. Software RAID over 3rd party SCSI controllers is plenty fast for most jobs... if its not we go buy RAID offload enclosures from Xyratec or somebody.
PERC... good hardware plus shitty Dell firmware. What were they thinking?
Wave Race fucking owned.
That game was so much fun to play... and all my friends try to race me and they're all like WTF HAX because all they do is sit around and play GTA all day. lmao
I must concur that Diet Mountain Dew has the most pleasant and consistent flavor of any diet drink (when compared to its sweet counterpart). Actually, I prefer the Diet variety to its full-sugar cousin, the sweetness is not as intense and it has a more pleasant citrus flavor. Very little aftertaste, essentially none. Alas very few places other than supermarkets and convience stores carry it. The only place that I know of that has it in fountain form is Quiznos (which is a good excuse to go there). :-(
Even more hard to find (and even better tasting) is Diet Mountain Dew Code Red.
They don't even stock the 20oz in 7-eleven anymore.
nt
This is true.
It doesn't necessarily make it less crash prone. But it does make it instrumentable if it proves to be unstable (you could easily trace, debug, intercept, or otherwise validate the requests the blob made if so needed).
Furthermore, the kernel mode portion would merely be relaying commands to trusted memory-mapped regions and IO space requested by the process initially (limited by configuration files, perhaps). Most kernel crashes are the cause of errors (pointer mistakes, buf overflow, race condition, etc.) in the complex driver code which "trap" the system in kernel space. The user space portion would likely instead SIG11 and die... if it left the hardware in a weird state it could be fixed by simply restarting the driver program which would, at its outset, send RESET type commands to the device putting it in a known state.
The largest problem I see is that it isn't possible to easily recast a userspace driver program into a device node without a mechanism like FUSE. It only works if the hardware target in question is nearly always accessed behind a userspace library (OpenGL, libalsa/libjack/OpenAL, libusb).
Most drivers don't need to run in kernel mode (read: any USB device driver)... or at least they don't need to run in response to system calls. /dev/dsp and /dev/input/mouse and such could be rewritten as kernel-threads that dispatch requests to and from other kernel threads servicing physical hardware in the system you can provide fault-isolation and state reconstruction in the face of crashes without incurring much overhead. Plus user processes could also drive these interfaces directly so user space programs could talk to hardware without needing to load in dangerous, untrusted kernel modules (esp. from closed-source hardware vendors).
The hardware manipulating parts kernel should stick to providing higher-level APIs for most bus and system protocols and provide async-io for kernel and user space. If most kernel mode drivers that power your typical
Or am I just crazy?
Yeah but microkernels seems like taking things to an extreme that can be accomplished with other means.
Dell and HP don't sell "propietary" service manuals to authorized service representatives.
Any document the person on the forum who would have cared to link to a supposedly incriminating page could have linked directly to Dell's or HP's support sites.
Nearly _every_ computer and computer component is manufactured/assembled in China. The exceptions are the large-chip manufacturing fabs run by big chip giants in Germany, Malaysia, India, and other semi-technologically savvy areas with cheap labor.
Rip your subtitles to a vobsubout .sub/.idx pair. .idx file, you can see what colors it intends to use on playback in the header file. You can change them there (they're just like HTML colors, #RRGGBB). .idx may be broken if you only get white text on playback. Check to make sure it's not trying to display a subtitle image at a weird time (often happens when ripping across cells when selecting chapter ranges)... that can confuse the playback software and cause it to miss settings or not display subtitles at the right times. I often have to hand edit the .idx files and tweak them myself.
If you check the
Your
Then again it could be an artifact of encoding the subtitles into the resulting avi. I don't use that method anymore, ever.
I encourage you to not encode the subtitles in the movie itself... but bundle it along (either as seperate files or as part of an OGM). It allows the end user to change color/transparency, move it, display it elsewhere (in the black bars going from widescreen->full, for example), and it improves encoding performance.
transcode (and then also mjpegtools) are kinda more of the attitude: do exactly this one thing and no more... for use in big conversion scripts and the like.
mencoder is more like: I just want you to convert this into this with this restriction and I don't care how you do it.
Of course, sometimes it guesses wrong about what you want it to do. Getting it do things a certain way can require black magic and animal sacrifice.
BSPlayer does not link or bundle in a full ffdshow library. It can leverage the ffdshow DirectShow filter to play a lot of media types without using other WM/DS libraries (people often prefer the features of ffdshow in MPEG2/MPEG4 over filters bundled with DVD drives and/or DivX). Usually you find BSPlayer and FFDShow bundled together, for example, in the KLite Codec Pack.
:-/
However, BSPlayer is a much better parser of video container formats (ASF, WMV, AVI, OGM) and MPEG transport streams than most other players out there (maybe with the exception of VLC). All of them are better than any versions Windows Media Player.
So it can handle broken, badly indexed, or partially downloaded files with ease.
Additionally, like VLC, mplayer and MPC, it can handle extended features in video containers that many other players (Windows Media Player included) omit. For example, multiple video streams, subtitles, multiple audio tracks, etc.
General Purpose complex CPU is slower than Single Purpose parallel CPU at the operations the Single Purpose CPU is designed for!!! Film at 11
I mean are you fucking kidding me? Banias, Dothan, and Sledgehammer do the GPP thing just fine at the same power/performance level as PPC or USPARCIV.
Of course NVidia would love to be put out their own CPU. Because then they could control THE WHOLE SOLUTION! CPU, chipset, video and sound! All they'd need is a HD manufacturer to partner with...
It's not the x86 dragging them down, its having to make their chips compatiable with the products of the rest of the industry. Big friggin deal... welcome to the modern tech marketplace.
No modern webbrowser allows cookies to "leave a trail" of where you've been anymore.
Cookies are only sent in requests to the domains that created them in the first place.
I guess one valid point would be that ad-networks can sort of track you by determining what sites who use their banners you visit based on referrals headers. The issue there is:
1) There's no user-identifiable information that links your cookie to the pages you visited feature that adnetwork's banners aside from your IP address at the time
2) Adnetworks don't care _who_ you are. They use the cookie to determine what sites should be considered "related" so they can better determine how to distribute banners. (That is, banners that do well on one site might do well on sites visited by people who visited the first site)
3) You can always block referers to images/iframes hosted on external domains (adnetwork stuff). IIRC this is the default now on Firefox and Safari... so... its irrelevant now.
I dunno about it being a Linux "litmus-test"
/usr/ucb in the front of your PATH.
ps -ef vs ps aux is just a SysV vs. BSD thing.
Of course on Solaris the aux thing might make sense if you had
But I learned on Solaris and I use ps -ef. Which is the way-we-do-things-now (TM).
They make lots of stuff.
* EPIA small/home theatre CPUs and mini-ITX boards
* AMD and Intel chipsets (K8Tx00, PT8x0, CX700, etc.)
* basically any southbridge not made by NVidia or Intel
* Envy24 audio chips (lots of cards, from Hercules to Terratec to M-Audio)
* VIA Velocity GBE and Rhine 10/100E ethernet chips (used in many, many desktop cards)
* A metric fuckton of small desktop switch integrated platforms... and a bevy of wireless MACs
Plus the S3 video chipsets are used in a whole lot of video-on-board motherboards even if the north/southbridges AREN'T from VIA. Basically if your server doesn't have an ATI Rage-XL or Intel IGP on it, then its got an S3. Tons of HP and IBM rackmount systems have them.
(awhile back)
Unix systems allocate a fixed area in each "cylinder group" to store the inode tables. Which means it doesn't get fragmented. OTH you can run out of inodes (or, conversely, you can give it too many and waste potential file storage area). NT has a dynamic allocation scheme but since the MFT is a file too it can get fragmented if you let your drive fill up.
That being said, probably the only file that matters performance-wise if its fragmented is the MFT on a Windows System. Fragmentation is not a big a deal as it seems... what's more of an issue is the availability of coalesced free space in which to write new data.
Unix systems don't have defragmentation tools because the allocation schemes are usually simpler and degrade nicely, and also because fragmentation for reads is pretty much irrelevant since on a multiuser system the drive head is going back and forth all over the disc anyway. Typically the system queues up all the pending IO requests and reorders and issues them in an elevator up-and-down pattern to maximize throughput.
If your unix system gets so fragmented that single-user performance is suffering and writing files takes to long, the "answer" is to tar up a subdirectory, or use find with a size>Xkb restriction, and copy them somewhere else, delete it, and then put it back on the volume. You can even do that to system files while the system is up since unix doesn't care when files get used/replace that are currently being used (unlike Windows). That is just as good as an in-place defragment.
1) Remote desktop == Remote assistance which means that you can help people with their spyware problems remotely without walking them through a Timbuktu or GoToMyPC install (blegh)
2) Volume Shadow Service == Restore Points, driver rollbacks, and backups of your system drive that actually work.
3) Those locking improvements are really important now that everyone has hyperthreading or dual core machines. It makes applications quite a bit snappier interactively, and helps when your doing stuff like burning to a CD while streaming a movie from another drive.
4) The 64-bit memory map stuff is very important because it removes a lot of stupid limitations that you ran into before. When you started seeing systems with > 1GB of memory the system wouldn't page it as effectively and the system was swapping a lot more. Also if you work with multimedia/content creation apps there were limitations to the file size you could use for memory mapped access (which is more useful when doing editing when you both seek and stream a lot... so developers would have it fall back to the shittier open/read/write-style access API).
And did I mention a built-in bluetooth and wireless stack?
These changes had a lot more implications in 5.1 over 5.0 than you give them credit for.
I find myself frustrated when going from a 2003/XP system to a 2000 system when things I expect to be there or for it to be able to do aren't there or are done in a less sane way.
once you strip it down. I don't bother and just install server 2003 (same thing really)
XP adds over 2000:
*) Volume Shadow Service (finally...)
*) MFT defragmentation support
*) 64-bit virtual memory support and 64-bit file mmaps
*) application prefetch
*) better page retiring algorithms and PTE management
*) new (faster) locks for kernel objects for SMP
*) Built-in Terminal Server and detachable console (nee remote desktop)
*) support for read-only NTFS (like on a CD or ZIP disk)
*) redesigned registry and configuration subsys (much faster, less mem, mostly lockfree)
*) slow int 0x2e syscalls replaced by sysenter/sysexit (big difference in context switches)
and a bunch of other little things behind the scenes.
So its worth going to NT 5.1/2 if you can whip it into shape.
If you go to yours or another users's page, you can see how they tagged articles (and whether they replied). Presumably this is a quick way to get a feel for what a user thinks about articles.
I understand the ability to search for articles tagged a certain way is also forthcoming.