I don't know, maybe this falls under the category of hardware, and the article is primarily about software, but sometimes, when I'm working on someone else's computer trying to solve their endless problems, I really wish I had a sledgehammer.
It's simple, really.
Press 'escape' during POST, to enter the BIOS setup program. On the 'Advanced Options' page you will find a section called 'ports'. Change the value of the 'Parallel' field from 'Bi-Directional' to "ECP'.
Not many motherboards older than Pentium II class types had boot from cdrom capability.
One of us has a terrible memory! At the end of 1994, I ran a server lab with 25 Pentium 60Mhz and 2 Pentium 90MHz. These were old enough to have the maths bug in the FPU.
Everyone of these machines booted from a 4x or 8x ATAPI CD-ROM drive.
My home machine was a glorious 486-66, with 48 MB RAM (!!!) and a Texcel 2x SCSI CD-ROM attached to an Adaptec 1542. I booted all the NT pre-releases in '93-'94, OS/2 2.1 and a huge number of Walnut Creek CD-ROM releases from '93-'95. This same kit booted on 386 - if you were so inclined.
It is absurd that these people are suffering undeserved deaths with our support and tax money - and most seem to be concerned that economic inconvenience will disturb their leisure and recreation.
We need a little lift, to feel good about ourselves. I mean, just because we are shooting the brains out of babies half-a-world away - Doesn't anyone else feel like they need a little regular old entertainment?
At least this thing hasn't put a crimp in the American idol schedule!
If Dvorak had even half a real clue about this, he would have mentioned OS X inheritance of OpenStep and OS 7's ability to run "Fat Binaries" - Multiple binary executable formats sharing a single resource-fork. ISV software was frequently shipped as single binary, that ran on NeXtStep, as well as Sparc and Intel OpenStep. Why no mention of this, or the Mac binaries that ran PPC/640xx ?
Also, he mis-understands Marklar. Apparently, this is a complete x86 Intel port of OS X. It acheives very little in targeting Itanium as a processor, as x86 is as much another slow emulation like PPC.
Yes, MS IE dones NOT respect MIME types, and MS has no fix here! I think IIS may well respect these - but URLScan is tasked to protect the clients too.
I do not know if the architecture is "broken". Many parts are implemented poorly, or with a bias to err on the side of user simplicity - without regard to systemic consequence. I agree that the net effect of these manifests the same behavior as a fundamental architectural deficiency.
I work in InfoSec. The longer I do this, the less bias I have towards one imlementation or another. You move your problem around the system - like a puzzle of sliding tiles. That said, I use *N?X systems as a matter of course, and don't have a Win box without Cygwin...
Exchange 2000 represents its DBMS message store as a filesystem, mounted as M: (!)
I think that Exch4/5 used a MAPI-style client-connection to get to the message store. In 2000, the M: drive is browsable -with the right perms- and full of sub-dirs named for each of the NetBIOS compatible logon names of the recipients. These appear to contain subdirs for mailbox folders, with messages as discreet filesystem objects. These are in the form [subject-line text].msg.
For message access, OWA constructs mailboxes by pointing to the.msg files as URLS - including ASCII/Unicode conversion to Hex for delimiter chars ( [subject-line%20text%2Emsg] ).
URLScan is pretty 'dumb'. It doesn't have complex rules, just a text-file config, with prohibited extensions (.exe,.com,.bat) and prohibited URL combos (%2E%2E%5C, %2E%65%78%65). I'm pretty sure that if ".." is prohibited, then the first two "dots" match the rule, URL is blocked, and a 404 is generated by IIS. You never parse farther down the URL. You will also block attachments with.exe extensions, because they are represented as URLs too. URLScan behaves just like you are trying to submit input to an.exe on the webhost, even though this is a simple GET, without a form.
The point is, this works most of the time. When it doesn't, you have a high-frustration situation for the user, with arelatively involved technical explaination. Do you want this situation when the user is a Senior Vice-President?
* IIS must be secured against cross-site and Unicode attacks. In reality, this means URLScan and IISLockdown. URLScan often makes undeliverable, messages which can be accessed via the Outlook 'fat' client. Example: the message with a subject-line 'This is the Visio...' will be acceptable to Outlook. OWA will turn this subject-line into the document name at the end of a URL. URLScan sees 'https://(fq.servername)/exchange/This is the Visio....msg', and parses the sequence of four 'dots' as a possible directory traversal. Access is denied! User sees a 404, big PITA. Expect lots of tech support calls on issues similar to this one.
* All the groovy advanced features are supported only under IE. Other browsers get a functional, if unexceptional subset. There is no activeX plugin or anything - MS just uses nifty, DHTML and VBScript for drag-n-drop, etc. in OWA. The server-side ASP on OWA effectively generates a different, alternate interface for non-IE clients.
Weigh your options, and see if it isn't better to publish Exchange access through an SSL-style VPN appliance like Neoteris or Aventail.
If you were a user of MacOS X, you would understand what a second-class citizen this has been in terms of Java. Native GUI support (Cocoa, Quartz) was non-existant. Java ran great - with CLI! This was a REAL shortcoming for users, untill now.
With a fully accelerated GUI and enterprise security features, OS X cannot be summarily dismissed as a client OS by business, at least not on the JVM features.
Proprietary Unix isn't a magnet for digruntled former OS/2 and BeOS users... Sorry, Nick P!
Honestly. Vendor Unix is usually tightly coupled to hardware. The advanced features often come out of this coupling, and are hardware specific. Linux does need a generic and abstract logical volume management system - the Sistina LVM is about the level of Solaris DiskSuite, minus GUI. IBM is implementing a superior and backwards compatible system: EVMS. If this makes it into 2.6, Linux will equal HP/UX and AIX here. There is also NUMA and ccNUMA work going on. This will kick a**.
Slashdot Mirror
Or is it really fourth?
Howabout:
'Self-absorbed navel-gazers' or 'Digital Sanitation Engineers'.
Systems Administrator is FINE! And I hope to god they don't try to change the name of SysAdmin Magazine.
It's simple, really.
Press 'escape' during POST, to enter the BIOS setup program. On the 'Advanced Options' page you will find a section called 'ports'. Change the value of the 'Parallel' field from 'Bi-Directional' to "ECP'.
Reboot, and your friend can now print!
=-)
One of us has a terrible memory! At the end of 1994, I ran a server lab with 25 Pentium 60Mhz and 2 Pentium 90MHz. These were old enough to have the maths bug in the FPU.
Everyone of these machines booted from a 4x or 8x ATAPI CD-ROM drive.
My home machine was a glorious 486-66, with 48 MB RAM (!!!) and a Texcel 2x SCSI CD-ROM attached to an Adaptec 1542. I booted all the NT pre-releases in '93-'94, OS/2 2.1 and a huge number of Walnut Creek CD-ROM releases from '93-'95. This same kit booted on 386 - if you were so inclined.
It appears that these files were made available on IRC, and are being subsequently transferred to ISO images of 2003 EE.
I leave it to the reader to figure just which USENET group carries this traffic.
MS needs to 'scour the Internet' for these people? They run one of the Websites in the C-Net article!
Just....KNOPPIX!
A picture is worth a thousand words.
It is absurd that these people are suffering undeserved deaths with our support and tax money - and most seem to be concerned that economic inconvenience will disturb their leisure and recreation.
At least this thing hasn't put a crimp in the American idol schedule!
I think this could help with the FreeNet 'speed' issue. Is anyone listening here?
: thie3e3 mijitz hayufv awul gann
Gone and all but forgotten!
: ownlee I, daddyo, remain--to deue bad tel wiyiuth thiee efil woneill.
And based on past battles, you couldn't fight your way out of a midget packing box sans styrofoam peanuts!
--
Bill O'Neill (woneill@pobox.com)
Toynbee ideas in Kubrik's 2001
http://www.pobox.com/~woneill Resurrect dead on planet Jupiter
Do they have a custom Linux distribution for this robot? I have a great name for this! HOAX
She was Sporty Spice's sister - the short-lived replacement for Ginger Spice.
Also, he mis-understands Marklar. Apparently, this is a complete x86 Intel port of OS X. It acheives very little in targeting Itanium as a processor, as x86 is as much another slow emulation like PPC.
Oceania has always been at war with Eurasia..
BTW, Iran regularly sees better popular turnout at the polls than the last 8 U.S. Presidential elections.
He won in 2000, after all...
Yes, MS IE dones NOT respect MIME types, and MS has no fix here! I think IIS may well respect these - but URLScan is tasked to protect the clients too.
I do not know if the architecture is "broken". Many parts are implemented poorly, or with a bias to err on the side of user simplicity - without regard to systemic consequence. I agree that the net effect of these manifests the same behavior as a fundamental architectural deficiency.
I work in InfoSec. The longer I do this, the less bias I have towards one imlementation or another. You move your problem around the system - like a puzzle of sliding tiles. That said, I use *N?X systems as a matter of course, and don't have a Win box without Cygwin...
I think that Exch4/5 used a MAPI-style client-connection to get to the message store. In 2000, the M: drive is browsable -with the right perms- and full of sub-dirs named for each of the NetBIOS compatible logon names of the recipients. These appear to contain subdirs for mailbox folders, with messages as discreet filesystem objects. These are in the form [subject-line text].msg.
For message access, OWA constructs mailboxes by pointing to the .msg files as URLS - including ASCII/Unicode conversion to Hex for delimiter chars ( [subject-line%20text%2Emsg] ).
URLScan is pretty 'dumb'. It doesn't have complex rules, just a text-file config, with prohibited extensions (.exe, .com, .bat) and prohibited URL combos (%2E%2E%5C, %2E%65%78%65). I'm pretty sure that if ".." is prohibited, then the first two "dots" match the rule, URL is blocked, and a 404 is generated by IIS. You never parse farther down the URL. You will also block attachments with .exe extensions, because they are represented as URLs too. URLScan behaves just like you are trying to submit input to an .exe on the webhost, even though this is a simple GET, without a form.
The point is, this works most of the time. When it doesn't, you have a high-frustration situation for the user, with arelatively involved technical explaination. Do you want this situation when the user is a Senior Vice-President?
* IIS must be secured against cross-site and Unicode attacks. In reality, this means URLScan and IISLockdown. URLScan often makes undeliverable, messages which can be accessed via the Outlook 'fat' client. Example: the message with a subject-line 'This is the Visio...' will be acceptable to Outlook. OWA will turn this subject-line into the document name at the end of a URL. URLScan sees 'https://(fq.servername)/exchange/This is the Visio....msg', and parses the sequence of four 'dots' as a possible directory traversal. Access is denied! User sees a 404, big PITA. Expect lots of tech support calls on issues similar to this one.
* All the groovy advanced features are supported only under IE. Other browsers get a functional, if unexceptional subset. There is no activeX plugin or anything - MS just uses nifty, DHTML and VBScript for drag-n-drop, etc. in OWA. The server-side ASP on OWA effectively generates a different, alternate interface for non-IE clients.
Weigh your options, and see if it isn't better to publish Exchange access through an SSL-style VPN appliance like Neoteris or Aventail.
I like your exteded similies on the naturalization status of Java in various foreign territories! Funny!
'Course, I meant OS X as a first-class citizen in Java-land - your inversion is more revealing of the situation.
Read the announcement carefully.
If you were a user of MacOS X, you would understand what a second-class citizen this has been in terms of Java. Native GUI support (Cocoa, Quartz) was non-existant. Java ran great - with CLI! This was a REAL shortcoming for users, untill now.
With a fully accelerated GUI and enterprise security features, OS X cannot be summarily dismissed as a client OS by business, at least not on the JVM features.
Honestly. Vendor Unix is usually tightly coupled to hardware. The advanced features often come out of this coupling, and are hardware specific. Linux does need a generic and abstract logical volume management system - the Sistina LVM is about the level of Solaris DiskSuite, minus GUI. IBM is implementing a superior and backwards compatible system: EVMS. If this makes it into 2.6, Linux will equal HP/UX and AIX here. There is also NUMA and ccNUMA work going on. This will kick a**.