It would be nice to code everything from scratch because of the control you have over bugs and performance, but it's not always very practical. The reason you use a black box library in the first place is so that you can save time by not having to reimplement all of it's functionality, which allows you to focus on coding the core functionality of your program (instead of first having to recreate a bunch of base classes). The library's developers also may have found solutions and/or optimizations which you may not have considered (because that functionality is their whole focus), they've already put the time into it's development so you don't have to.

Hmm...UNIX...the same folks who origionally included the passwords in the passwd file which is readable by all users on the system. It doesn't mean that UNIX is shit. Like everything in the computer world they didn't plan for exploitation and had to learn a valuable lesson before the design was updated (ie. passwords are now stored in the separate 'shadow' file which is not readable by all users).

NTLM was badly designed and was replaced by Kerebos encryption way back in Windows 2000. I think Microsoft might have learned a bit about securily hashing passwords in the almost two decades since NTLM was designed.

No, Microsoft really doesn't give a sh*t about which OS is used to host a bunch of DNS servers, and they don't give away Window Server licenses to any company because that's their biggest money maker: selling their server products to corporations.

MS could care less about home users pirating a $130 copy of Windows 7. The real money is in selling $1000+ server licenses to companies for many servers, as well as having to buy things like seat licenses and other expensive server products (like SharePoint). Plus those companies also have to purchase "professional" level Windows clients for their workstations which connect to their servers. That's why they regularily audit corporations and organizations for license compliance in person, and invent schemes like WGA Validation to handle the home users who they don't have time to care about.

Yeah I've had to do Debian installs in 4MB of RAM and it sucks. You needed a special boot floppy with a custom built kernel and minimal install interface that can fit into that 4MB of space effectively. It didn't run too badly but I had a larger hard disk (and thus a lot of swap space as well). I couldn't imagine having to also deal with a 40MB hard disk!

But I agree, there is a huge leap in performance when you move from 4MB to 8MB of RAM and you don't have to fiddle around with things as much.

You may also find it interesting that Windows 95 installed a specially built kernel if you had 4MB of RAM. If you had more than 4MB it copied over the regular version. Apparently Microsoft also found it difficult to get Windows to work in 4MB of RAM, but 8MB was no problem at all.

You should try and install smartdrv.exe (the MS-DOS disk cache driver) onto the boot floppy. I don't know 100% for sure if the Windows 95 installation is affected, but Windows NT 4.0 and Windows 2000 installations started from MS-DOS were incredibly slow to copy the files if smartdrv.exe wasn't loaded (the file copy phase of the installation would take at least three times as long).

Another thing that makes Windows 95 into a swap zombie is if the disk controllers are loaded in 16-bit compatabiliy mode. IIRC under origional Windows 95 (not as much under OSR2), a lot of laptops had this issue because it didn't include many chipset drivers. Even common Intel chipsets will have performance issues until you install the proper chipset software.

Of course it could also be caused by a variety of hardware related issues...

For example, I've installed Windows 95 on 486's with 8MB of RAM and had decent performance, yet an IBM Aptiva with a Pentium II and 128MB of RAM struggles and seems to perform worse. The Aptiva used an ALi IDE chipset which was really slow and IBM paired it with a Quantum Bigfoot hard drive which was also really slow due to it's large platter size.

Apple are using the courts to shut down competitors, which means that they do not back their own capacity to compete.

Umm...Apple has recently moved past Exxon Mobil on the list of most valuable companies in the world. In the mobile arena they've been able to kill off Windows Mobile, Palm and even RIM is suffering lately. Apple knows they can compete in the mobile marketplace.

Apple also knows that they need to defend their patents when issues arise or risk losing the ability to defend them later on. Look at the shitload of Apple technologies that Microsoft stole/copied in the earlier days of Windows and ask yourself why Apple might be a bit aggressive when defending their patents.

Sorry fanboy. IBM had hyperlink before Apple existed. Haven't you ever used big iron?

No I don't use big iron IBM systems...and I suspect many people here don't either. So please tell us which part of their big iron technology used hyperlinks among different documents long before others, and when it was released. Just bragging because you've worked on big iron doesn't prove shit...

Hypercard was released in 1987 and was definately one of the earliest applications to use hyperlinks. It's not some kind of fanboy thing, even Wikipedia says this.

The bit length measurement of a CPU or application architecture can really be of any size, it is usually determined by the size of the address bus, registers or data bus. The most common archeitectures use lengths in powers of two but there are a few platforms which do use odd bit sizes (a lot of these were older systems like the PDP and the System/360). IIRC the 18-bit PDP system could perform direct manipulation of individual bits in it's registers, so having an odd bit length didn't matter. Newer archetictures often manipulate the entire length of the register so it makes it much easier and more efficient to keep sizes at a powers of 2.

Windowd 95 was also a fully 32-bit operating system, It booted into 16-bit MS-DOS initially but then entered full 32-bit protected mode when the kernel (krnl386.exe) was loaded. The 16-bit MS-DOS session is discarded and has no effect on the 32-bit part of the operating system when it boots into protected mode. They probably used MS-DOS to provide basic memory, hard disk and file system support for the purpose of locating and booting the kernel (similar to what NTLDR does on Windows NT), and also to provide full MS-DOS compatibilty by providing a way to boot into MS-DOS only.

So nothing freaky happens (like ... you use windows so that non respond in a standards-compliant manner)

If you're working for almost any large corporation you probably have Windows workstations and the easiest way to manage a large number of them is with Active Directory. They probably also will connect to Windows servers for shared file storage and printing, and they may even use other popular Microsoft products like Exchange and SharePoint.

In this environment Windows is the standard and it is the Linux box which is the outcast and cannot interoperate fully with the rest of the corporate network.

Don't try and spin any FUD here about corporations not using Microsoft either...I have worked for companies like IBM, who is a big time Linux supporter and has equivellent offerings for almost every line of corporate software Microsoft offers. You know what? They were running Windows workstations and had Windows servers for situations where they were simply the better solution.

If you really have any understanding of servers you would know the strengths and weaknesses of both Windows and Linux server implementations and be able to plan accordingly.

The people that seem to love Apple seem clueless about what Apple have actually done... closed of a mini handheld computer so only software they endorse, can run. It's like a protection racket in a modern day world. Developers pay and can have their app removed without notice - legally. They have to pay for the ability to make software, pay to have it hosted and pay for every sale they make. APPLE really are SHAMELESS

Anyone who's used Windows Mobile for any length of time will tell you how awesome it is to have applications kept under such tight control. Windows CE was absolute software freedom and left almost everything up to the developer or user, who are not always the most competent people when it comes down to it.

Windows Mobile allows applications to install anywhere and in a variety of fashions which are pretty much entirely up to the application developer. There is also no centralized source for downloading applications (like the app store) and you just get them from a variety of web sites. Apps can be written using a large selection of developer tools which span the many different releases of Windows CE (This includes languages like eMbedded Visual Basic which was completely awkward to use for many programming tasks, and the runtime had memory leaks and bugs which you had to be aware of and work around). There is no type of quality control either, you will find many applications that don't work, display incorrectly, are buggy or just plain cause the phone to reset.

I really like how apps in the iPhone app store always install and work properly just like that. You don't have to sit in suspense during the app install, and you aren't ever concerned about where the fuck tfiles and registry entries are being created. It's a really good win for software freedom when you have to waste time removing registry entries and files from obscure locations on your phone because the app developer could be a retard or just plain ignorant prick.

Yes I thought of Smart Tags too when I first read the details of the patent. IIRC Smart Tags were introduced in Office 2000 while the Apple patent is from the mid-nineties. Microsoft and Apple probably have a cross patent licensing deal that allowed Microsoft to release Smart Tags without worrying about the Apple patent anyway.

Having said that I do prefer the classic "bars" on the Windows 7 task bar, mainly because I have loads of windows open at once. In that mode there is quite a bit of space around the clock wasted because it remains the same width it was when the taller task bar allowed the date to be shown under the time.

I have noticed the same issue, the new taskbar interface and aero peek are great but fail hard when there are many windows open under a single process. I am especially frusterating when I use Internet Explorer because I often keep multiple IE windows open with many releated tabs open under each window. The classic taskbar would simply display each IE window as a separate icon (or list entry under taskbar grouping). Windows 7 displays each tab as a separate list entry under the IE taskbar icon, regardless of the window which the tab is contained in. The result is that you have to scroll through a massive list of open tabs.

When Apple did their origional research, they found that the menu bar in the same top location allowed users to locate it and it's commands more quickly (as they don't have to search for it). The GUI is no longer a new concept for users. Ater using Windows for a while, users also develop a technique where they pull back the mouse slightly and in that instant perform a quick visual check to locate the menu bar. The slight delay to find the menubar is no longer a large disadvantage to the length of time it takes to move the cursor to the top of the screen on the primary display,

I think the single menubar concept in OS X is a relic of the origional MacOS UI design. Under classic Mac systems the model works because most classic applications only display one document window which is just switched out when the user selects another open document in the application. Multitasking was also pretty simple and was not used like it is in today's operating systems. Under the old MacOS, the single menubar concept worked well because there was no real need to group many applications and document windows like there is under modern operating systems. There were also fewer systems with multiple monitors and they had much smaller resolutions to move the cursor over.

We all saw what happened when Microsoft removed the menus from Office 2007 and introduced the ribbon. The UI designers of OS X probably kept it to avoid confusion and problems with developers and users initially moving to OS X. Now they are stuck in a position where it would be a massive undertaking to remove it from existing OS X applications.

You could probably go to the flea market and buy an entire system with Windows 2000 for like $40. Or developers could spend hundreds of dollars making sure their web page will work on whatever ancient IE version you have on Windows 98.

Yeah my commodore 64's hardware is still working too. Even if I could somehow load a web browser from the casette drive, I wouldn't have the unrealistic expectation that web developers are gonna bend over backwards to support my crappy ancient system. Yeah that would be nice if we could watch youtube and check gmail on 20+ year old hardware, but it's rediculous to actually expect it.

Mobile operating systems don't have performance and resource counters for every process because they would provide too much overhead and they are trying to squeeze out every extra milisecond of battery life that they possibly can, They need to keep memory usage and processor time to a minimum, not only because they are limited in size, but also because using them draws battery power as well.

Also consider that most smart phone operating systems didn't have full multitasking capabilities until recently. There was no reason to track process CPU or memory usage because they pretty much had all of the processor time, available memory and network I/O exclusively while they were running.

We must always remember that a smartphone is not a computer. It shouldn't have a console and advanced system diagnostic tools. The system shouldn't need to monitor and report process resource usage. Why? Because a smart phone shouldn't be running that many background processes in the first place!

Desktop computers have magnetic hard disks which they can use for storing files and inactive memory pages. They have large amounts of memory for caching and buffers. They can constantly run many pieces of hardware without worrying as much about battery life and space in a tiny form factor.

I can understand that maybe you do need to have so many different applications running in the background to poll for facebook updates and whatever else. If that is the case then the OS vendor should look at developing a common interface which unifies and better controls them. For example, instead of having processes constantly running and opening TCP/IP connections to various servers to check for updates, Apple decided to implement push notifications.

If it's such a badly written application, it shouldn't be published in the app store. It should be available under a separate repository, or require manual installation. That way only people who want to hack around their phones will be able to install them. It should definately not be up to the average user to start worrying about things like that.