XML Registry? Maybe, one day. It's more likely than you think. They made the IIS Metabase an XML file for IIS 6. They're making application and site and user and machine configuration files for.Net as XML. XML as a configuration file format is becoming quite popular here in the land of the Evil Empire.
Sure they're evil - but they offer great benefits!
I can see why you consider this a problem - an oversight on the part of the CLI developement team, but you fail to follow through and explain just how this issue makes the CLI unsuitable.
The command line to which you refer is the command shell cmd.exe. It has less to do with dos than the Korn shell has to do with bash - at least bash and Korn are 32 bit apps....dos, for all the extenders (such as what came with my old Borland Pascal 7.0 IDE, I think it was PharLap) was never a 32 bit beast. Cmd.exe is a 32-bit NT console-mode shell...that shares much of the old dos syntax. As far as any emulation goes, that's just yet another client subsystem (you do know that NT is a client-server OS, right? Win32 is not the native API on NT, it is just another subsystem, like OS/2 or posix. To run on this parenthetical comment on a bit more, I'll even mention that Win32 is certainly the preferred "language" on NT, though with.Net you can count on that changing. I'll stop Karma whoring now.) called the NT Virtual DOS Machine, or NTVDM. The Windows-On-Windows (WOW) subsystem runs on top of the NTVDM to extend this emulation to 16-bit Windows applications. All 16-bit dos/windows apps run in a single isolated process space by default, unless configured differently. In any case, the command shell is no more the OS executive in NT than bash is in GNU/Linux. The kernel houses the executive. And I mean both NT and GNU/Linux here.
Yeah. According to the API docs, it works on W2k any edition, XP Pro and.Net Server. No go for NT, but I feel sorry for anyone still working in NT anyways.
Actually, if he's running Win2k server, he can use the Job Object API This API allows you to set per-process limits on cpu, memory, user mode execution time, min/max working set, processor affinity, thread priority, UI restrictions, and security restrictions. I believe that Win2k Datacenter Server comes with a Job Object MMC for creating Job Objects / adding processes to a Job Object.
Email each journal entry to yourself. Use an email provider that isn't connected with your business if you feel that gives you an added level of trustability. Email is admissable.
I have several products like than installed on my WinXP.554.lin machine and I ha43n34 noticed any degradatiosafdjhbsf of my TCP/IP stack. So,8-9,-09u off the FUD, eh? %
How many are documentation errors? Variable naming inconsistencies? Refactoring notes? If you wanna hate MS, fine, but you lend credibility to that which you oppose when you fight with FUD.
The next advance leading toward ActiveX controls was the proliferation of 32-bit technology. VBX could only be compiled for 16-bit applications. Because COM was 32-bit compatible, it was the ideal basis for a new control technology. Furthermore, controls developed using COM could be written in any language and hosted in applications developed in any language. OLE custom controls (OCX) were introduced and replaced VBX technology. The original OLE control specification required each component to implement a minimum of nine specified interfaces, with a total of 60 methods and seven optional interfaces. This detailed specification of well-defined interfaces made OLE controls easy to use but tiresome for control developers to create.
The final incident motivating the development of ActiveX controls was the arrival of the Internet, particularly the popularity of the Web. Initially, the size of the control was not a concern because RAM and hard disk space were becoming less expensive each year, but the prospect of downloading controls over a slow connection made size a concern. If users became frustrated because a page with a large component was taking too long to download, they could find Web content elsewhere. The ActiveX control specification was created as an answer to this problem.
ActiveX controls are required to support only one interface, IUnknown, and two API functions, DllRegisterServer and DllUnregisterServer. These two functions ensure that ActiveX controls are self-registering; that is, they can add and remove the necessary information for their use in the target system's registry.
Actually, ActiveX isnt so much a misnomer for COM - although it is commonly misused as such - as it is a specific COM technology for creating visual controls. ActiveX controls replaced the vastly more complicated (read: royal pain in the @ss) OLE controls with a much simpler interface model - even though MS themselves violated their own published spec for the vtable with VBs implimentation. Yes, the whole terminology scheme for OLE/COM/MTS/COM+/ActiveX did get very muddy. <Disclaimer>No, I ain't no steenkin VB coder</Disclaimer>
Re:ya, I couldn't talk till I saw that book
on
Design Patterns
·
· Score: 2
Re:ya, I couldn't talk till I saw that book
on
Design Patterns
·
· Score: 3, Informative
Sure it's trivial. Much like the "trivial" civil engineering concept that a bridge built with too little concrete will fall down.
Certainly, anyone who does "real" (as opposed to integer?) development will have discovered many, if not all of these patterns. Why wait years, when you can read about them? Patterns are studied in architecture - and humans have been building structures far far longer than they have been crafting software.
Sure it's a catalog. It's also a lexicon. Does that invalidate it's utility? This is the kind of work that will help push software development into a estimable science instead of the craft it largely is today.
Java(tm) and C++ and are excellent ways to document implementations of systems. However, a design pattern is intended to be an abstraction that can be applied to ANY language. That way, a Smalltalk developer can say "Strategy" to a Java(tm) coder and they both share the appropriate context.
You seem rather hostile about this book, resorting even to ad hominim attacks upon the GoF, who have separately and collectively done a great deal to advance OOP. Perhaps you regret not having written down these obscenely trivial ideas yourself and peddling your own brand of software snake oil.
hmmm....I dont know. I do know that Air Force One does have a large array of different comm systems in place, including UHF SatCom and I believe Iridium-style gear. But to call it "AF1's broadcast system" sounds a lot like you're saying it has only one communications system, which is certainly not the case. AF1 has to communicate with NORAD, the Pentagon, the White House, not to mention being capable of patching into the conventional phone network.
I'm curious....which network library were you using, in both instances? I exclusively use DBMSSOCN, which is the TCP/IP library. I have (highly multithreaded, batch-style, network bound) applications which get a 10~15% boost from running local on the sql sever. I seem to remember that named pipes, which in many cases is the default library, performs poorly. Also, the OS version would have a lot to do with it, as TCP/IP performance has made drastic inroads since the days of NT4.
Hehe. We have several large AS/400s/iSeries and more RS/6000s/pSeries servers. They started sucking on the Wintel front, though, and now we're down to less than 5% of our Windows Servers being IBM. We also dropped IBM for Compaq on the desktop (6000+ machines). They are FAR more accomodating on the midrange stuff. They never really seemed to care when we dropped them.
It completely depends on the load the machine will experience. Maybe you could explain a little more what kind of traffic the server will be handling.
Here at work, I deal with SQL Server machines from 2x PII Xeon 450 all the way up to 8x PIII Xeon 733. For the loads we experience, more processors is better than faster/fewer, but I've not worked with P4 Xeons. Hyperthreading might actually tip that in your favor. Can you buy a 4x capable chassis with 2 CPUs? Maybe leave yourself some room. We do this here quite a bit - but we use Compaq Proliants.
Slashdot Mirror
TextPad.
XML Registry? Maybe, one day. It's more likely than you think. They made the IIS Metabase an XML file for IIS 6. They're making application and site and user and machine configuration files for .Net as XML. XML as a configuration file format is becoming quite popular here in the land of the Evil Empire.
Sure they're evil - but they offer great benefits!
Keep in mind that a major (if not the biggest) military user of satellite bandwidth is the navy - and dish size isn't constrained much on ships.
I realize this is offtopic, but just go get an ADAT off of ebay. They're 8 track digital, they can scale out, and they use VHS tapes.
I can see why you consider this a problem - an oversight on the part of the CLI developement team, but you fail to follow through and explain just how this issue makes the CLI unsuitable.
Heaven forbid you use some crappy old revision control system. That would make the current code set readable and legacy code retrievable. Ewwww.
The command line to which you refer is the command shell cmd.exe. It has less to do with dos than the Korn shell has to do with bash - at least bash and Korn are 32 bit apps....dos, for all the extenders (such as what came with my old Borland Pascal 7.0 IDE, I think it was PharLap) was never a 32 bit beast. Cmd.exe is a 32-bit NT console-mode shell...that shares much of the old dos syntax. As far as any emulation goes, that's just yet another client subsystem (you do know that NT is a client-server OS, right? Win32 is not the native API on NT, it is just another subsystem, like OS/2 or posix. To run on this parenthetical comment on a bit more, I'll even mention that Win32 is certainly the preferred "language" on NT, though with .Net you can count on that changing. I'll stop Karma whoring now.) called the NT Virtual DOS Machine, or NTVDM. The Windows-On-Windows (WOW) subsystem runs on top of the NTVDM to extend this emulation to 16-bit Windows applications. All 16-bit dos/windows apps run in a single isolated process space by default, unless configured differently.
In any case, the command shell is no more the OS executive in NT than bash is in GNU/Linux. The kernel houses the executive. And I mean both NT and GNU/Linux here.
Yeah. According to the API docs, it works on W2k any edition, XP Pro and .Net Server. No go for NT, but I feel sorry for anyone still working in NT anyways.
Actually, if he's running Win2k server, he can use the Job Object API
This API allows you to set per-process limits on cpu, memory, user mode execution time, min/max working set, processor affinity, thread priority, UI restrictions, and security restrictions.
I believe that Win2k Datacenter Server comes with a Job Object MMC for creating Job Objects / adding processes to a Job Object.
Isn't that redundant? Prophecy in steps 2 and 4???
Oh yeah...PGP might be a good thing here.
Email each journal entry to yourself. Use an email provider that isn't connected with your business if you feel that gives you an added level of trustability. Email is admissable.
I have several products like than installed on my WinXP .554.lin machine and I ha43n34 noticed any degradatiosafdjhbsf of my TCP/IP stack. So ,8-9,-09u off the FUD, eh? %
Cliched? Perhaps. Prior Art? Certainly. Matrix-like? Yes.
So what? If you don't want to go see it, don't. All sorts of painters painted flowers, after all. Its all been done.
Sure looks worthwhile to me, though.
63k bugs huh? Wow. Kinda like this
How many are documentation errors? Variable naming inconsistencies? Refactoring notes? If you wanna hate MS, fine, but you lend credibility to that which you oppose when you fight with FUD.
Actually, ActiveX isnt so much a misnomer for COM - although it is commonly misused as such - as it is a specific COM technology for creating visual controls. ActiveX controls replaced the vastly more complicated (read: royal pain in the @ss) OLE controls with a much simpler interface model - even though MS themselves violated their own published spec for the vtable with VBs implimentation. Yes, the whole terminology scheme for OLE/COM/MTS/COM+/ActiveX did get very muddy.
<Disclaimer>No, I ain't no steenkin VB coder</Disclaimer>
Tell that to Kent Beck
Sure it's trivial. Much like the "trivial" civil engineering concept that a bridge built with too little concrete will fall down.
Certainly, anyone who does "real" (as opposed to integer?) development will have discovered many, if not all of these patterns. Why wait years, when you can read about them? Patterns are studied in architecture - and humans have been building structures far far longer than they have been crafting software.
Sure it's a catalog. It's also a lexicon. Does that invalidate it's utility? This is the kind of work that will help push software development into a estimable science instead of the craft it largely is today.
Java(tm) and C++ and are excellent ways to document implementations of systems. However, a design pattern is intended to be an abstraction that can be applied to ANY language. That way, a Smalltalk developer can say "Strategy" to a Java(tm) coder and they both share the appropriate context.
You seem rather hostile about this book, resorting even to ad hominim attacks upon the GoF, who have separately and collectively done a great deal to advance OOP. Perhaps you regret not having written down these obscenely trivial ideas yourself and peddling your own brand of software snake oil.
hmmm....I dont know. I do know that Air Force One does have a large array of different comm systems in place, including UHF SatCom and I believe Iridium-style gear. But to call it "AF1's broadcast system" sounds a lot like you're saying it has only one communications system, which is certainly not the case. AF1 has to communicate with NORAD, the Pentagon, the White House, not to mention being capable of patching into the conventional phone network.
I'm curious....which network library were you using, in both instances? I exclusively use DBMSSOCN, which is the TCP/IP library. I have (highly multithreaded, batch-style, network bound) applications which get a 10~15% boost from running local on the sql sever. I seem to remember that named pipes, which in many cases is the default library, performs poorly. Also, the OS version would have a lot to do with it, as TCP/IP performance has made drastic inroads since the days of NT4.
Hehe. We have several large AS/400s/iSeries and more RS/6000s/pSeries servers. They started sucking on the Wintel front, though, and now we're down to less than 5% of our Windows Servers being IBM. We also dropped IBM for Compaq on the desktop (6000+ machines). They are FAR more accomodating on the midrange stuff. They never really seemed to care when we dropped them.
The only place it would matter is my big servers...and they're 8 ways.
For one customer buying one lousy server? I'd be suprised if they care....
I know Compaq publishes white papers on MSSQL best practices/tuning/sizing. Perhaps IBM does as well.
It completely depends on the load the machine will experience. Maybe you could explain a little more what kind of traffic the server will be handling.
Here at work, I deal with SQL Server machines from 2x PII Xeon 450 all the way up to 8x PIII Xeon 733. For the loads we experience, more processors is better than faster/fewer, but I've not worked with P4 Xeons. Hyperthreading might actually tip that in your favor.
Can you buy a 4x capable chassis with 2 CPUs? Maybe leave yourself some room. We do this here quite a bit - but we use Compaq Proliants.