Slashdot Mirror
Windows 95 Turns 10
ColdGrits writes "It's hard to believe it, but 10 short years ago today saw the launch of Windows '95.
Here is an archive of the Washington Post's story on the day. As part of the launch, Microsoft paid $12,000,000 for the rights to use the Rolling Stones' song "Start Me Up" (containing the prophetic line 'You make a grown man cry'). "
The Slashdot Project was our last, best hope for unbiased news.
It failed.
But in the year of the Linux War, it became something greater: our last, best hope - for blinding stupidity.
Grow up Taco...Windowsz?
While I'm sure many people here will make jokes about Windows 95, it was quite a leap in stability and usablility from windows 3.1. I don't think windows has had such an upgrade since then, nor do I think Vista will be that much of an improvment over XP/2000.
Actually, that quote should read...
Microsoft is delivering the same features they developed for us seven years ago.
Who do you think WROTE OS/2?!?
Besides, NT was already out and gaining popularity during this timeframe.
...Windows 95, despite all it's many flaws, was a lot of fun. It was stupidity to use it in a situation requiring stability, but as a gaming platform and all around PC OS, it was great to have at the time. Especially with the freeware that became rapidly available, it was a big laboratory for computer users. Remember, MS didn't have an app for everything back then, so if you needed one, you bought it or sought it out on the freeware sites. Though I'd used Unix in school, my first exposure to IRC was on Win 95, and I relied on the freeware IRC clients to learn. Same with the utilities and such.
I'd never owned an Apple, so I can't speak to what it was like to use one back then (were they using, what, system 6 at the time? I don't remember...), but while XP is more reliable, and I get a tremedous sense of "do it yourself" satisfaction with Linux (my primary laptop OS), I don't think I'll ever have as much pure fun as I did playing around with Win 95 when it first came out, warts and all.
Life is hard, and the world is cruel
In retrospect, it is amazing how long it took Microsoft to come out with those features relative to their competition. Inovators that they are.
Except for ending slavery, the Nazis, communism, & securing American independence, war has never solved anything.
While Windows 95 may have reduced the GDI resources problem of Windows 3.x, I hardly think anyone could credibly claim that Mac OS had good memory management before OS X.
Is it really correct to keep counting age after it's dead? When was the last time Windows 95 was available to buy? When was the last one someone actually used it(I'mnot counting museums here)?
I'd rather be flying
You're not the only one... I'm pretty sure I spent more time reinstalling Windows 95 then I actually spent on Windows 95.
Hell, my win98 box is pretty much open to any script kiddie out there because MS refuses to issue patches. "It's seven years old, Grandpa. Get with the times!"
My Chevy is a 1988 and, like my copy of Win 98, runs fine. I had to have the fuel pump replaced last year and I still have to change the oil, etc, but it gets me where I want to go.
Now, if a large defect became apparent (say, it exploded on impact like a Pinto or a Crown Vic), Chevy would recall it, even though it's 22 years old. If they didn't, the government would force them to.
Yet the government won't force Microsoft to patch security holes in win 98, let alone 95.
What's wrong with this picture?
You seem to have pretty rose coloured glasses for Win95. You talk about it like you used it in 1999, not 1995. Let me refresh your memory!
:)
:-D
Win95 was terrible for games. None of my games worked with it. None! Not until DirectX 5 and 6 could DirectX be said to have matured enough for general use. Nothing really good came out until then, either. Quake was still something you'd "Exit into DOS mode" for.
As for the Freeware, most of it was dreesed up Win32s apps or NT apps now able to be run (thanks to Win95 implementing full Win32). The MS Plus pack was a good example of the sillyness of the era: IE 1.0 came with it. That thing sucked. People were desperate for uninstallers that wouldn't hose the system (cleansweep, etc, came out around then). And the memory managers for DOS still sucked -- keeping QEMM 7 around was much better than using DOS 7's emm386/himem.sys!
If you had 16mb of RAM, Win95 was noticably bitchy compared to Win3.1. You needed at least 32mb of RAM, and at least a Pentium 120 to really have it go decently. That was a top-of-the-line computer until fall 1996.
Thankfully, Netscape 1.x was available and 32-bit then. Plus you could run it just as easily on an Indy or DECStation or Linux
The best thing about Win95 was that it included its own 32-bit Winsock implementation.
PS: System 7 came out in 1990! By the time Win95 was out, it'd been updated to 7.5ish (7.5.1 came out in March, 1995; 7.5.2 in August, 1995). This was a pretty decent OS for not having real guts to it -- Quicktime, Applescript, PowerPC support (for the "new" PowerPC CPUs), Powertalk, and easy to add/remove TTFs. Windows just barely got the TTF part with Win95. Windows Media Player in Win95 didn't come close to Quicktime!
Mock mock mock mock mock mock mock
--
Internet Explorer (n): Another bug -- that is, a feature that can't be turned off -- in Windows.
Oh come one. Sure, there are lots and lots of problems with Windows, but scripting objects is not one of them.
Ever try to work with filenames containing spaces? Ever need to manipulate data that represents a graph or tree (other than a directory tree)? Ever need to manipulate a bunch of spreadsheets (including layout)?
I've done each of those in bash and in WSH and I infinitely prefer the latter.
Using plaintext when possible is a great idea that I support 100%, but for some things it just plain sucks. And as soon as piping objects is made easy (as MS claims to be doing with Monad), objects will become more desirable still.
Honestly, the *nix world is rediculously smug when it comes to these things. For ages scripting was way better under *nix, but in the past years it seems that MS is where all the progress is being made. They're still not entirely there, but they're gaining ground fast.
Athletic Scholarships to universities make as much sense as academic scholarships to sports teams.
Sounds a bit overcomplicated to me, really. At least with the error prone parsing through text piped through app after app, I'm at any point able to thow a tee in the script and send the output somewhere that I can visibly read it and interpret it. I can also take that output and modify it slightly and send it manually back through the next step in the chain to do some additional testing. I'm not sure that simply examining the properties of the .NET object affords me the flexibility.
I'd also point out that I personally disagree with a lot of this obsession over object oriented code in everything these days. In a short script with a defined start and end, there's no need for the obfuscation of object orientation. I hate it when I see a huge generic class included by default on every page of a web application, even though some pages may only use 1 (or even NONE) of the functions within that class. At that point it's just a bunch of uneccessary overhead. It begins to seem like developers get use to that style of $this->crap and they can't get out of it
``Token Ring LAN''
...'' Those calls were a piece of cake.
Heh, I remember when some idiot at the brokerage firm I worked at tried to install Windows 95 on the Broken Ring network the weak after 95's release. It took down the whole network. He got escorted out on the spot. Long story short: Broken Ring was a nightmare to support regardless of the OS on the clients.
Later, when I worked at a Help Desk, I loved our clients that ran OS/2. For the most part, they never called. When they did call, the problem was almost always a training issue, ``Yeah, it works that way in Windows, but in OS/2, you have to
Supporting Windows 95 was a mixed bad. For the first two years after release, supporting the OEM version was a nightmare. To this day, I'm convinced that the original OEM version of Windows 95 was nothing more than an expanded beta test. The retail version, however, wasn't bad to work with at all. At least not at the time of release.
Of course, a couple of years down the road with OEM SR2, the OEM version of Windows 95 became vastly superior to the retail version. At that time, OEM SR2 was the best, easiest to support, Microsoft operating system ever. IMO, its reign as the King of the MS operating systems lasted until 98SE. NT 3.x and 4 were fiendishly difficult to support, mostly because of hardware incompatibility.
Well, one question I asked myself while struggling with Win95 in 1995 was "Will I feel the same nostalgia for this in ten years that I feel for homecomputers from the 1980s?"
I can finally say: "Not a single bit! I am glad it's dead. And I am looking forward for it's brothers to die too."
...isn't this called AppleScript?
And yes it is innovative, just don't mistake it for a Microsoft innovation. (should also probably give props to BeOS messages as well...)
Hmmph. I can never understand how Linux zealots are so enamoured with cryptic command-line tools. Man pages are pretty-much opaque, and require a Man page themselves to understand. GUI materials are self-documenting - you can see what you can do with them just by looking at them. Other platforms have perfectly intelligent methods for scripting GUI objects - it's nothing inherently flawed in the paradigm.
Plus, once again, buffer overruns are a function of a particular bad implementation of programming, not OOP in general.
Personally, I think the platform I can do the best scripting in is Python. Easy, sensible help system, good tools, nice syntax, etc. But also consider things like LabView, that can make a perfectly functional programming language and GUI-and-program system just by wiring diagrams together. Apple apparently has some goregeous innovations coming in the world of user-scripting.
But meanwhile most Unix nuts are still convinced that Bash is the be-all and end-all, despite having utterly bizarre gotchas (like the recent story where someone described how having a file called -r can result in rm * having the very unexpected sideeffect of deleting recursively).
Learning to do a new task in a pure-text environment is like trying to learn how to spell a word with a dictionary - you can't look it up until you know how to spell it. Likewise, you have no idea what tool you use for a task until you already know what that tool does, and then you have to read confusing documentation of how to use it. Meanwhile, a nice GUI lets you figure it all out just from checking out the widgets.
Unfortunately, just because _one_ company decides to leave it's GUIs without any coherent standard for scripted GUI access, all most other guis make this same omission.
All I know is that the win2k "find" screen makes 10x more sense than the grep command.
I can never understand how Linux zealots are so enamoured with cryptic command-line tools.
.NET that did things like statistical analysis of characters in large text files, because the input was minimal and it took less time than making a GUI. For the analysis of the actual encrypted text though, I wanted a GUI because it let me make changes in the decryption options and see the changes update across the screen, rather than comparing two text files of output from a command line tool.
The *ix command line is what I miss most when I use Windows systems (which is most of the time, currently).
It takes a little getting used to, but it lets you do all the things you *think* you should be able to.
For example, using tr I can replace characters or strings in a file or text stream as part of a batch process. On Windows I'd have to write a script or a program to do that.
Another *huge* benefit is that you can do massive batch processes without depending on a GUI app supporting it. If I have a command line tool that converts TIFF -> PNG or whatever, I can do tiff2png *.tiff *.png and be done with it. Some GUI apps like Photoshop might be able to do the same thing, but it would take more time to set up, and I may not have an app with that capability.
Man pages are pretty-much opaque, and require a Man page themselves to understand.
That I'll agree with you on. I've never been fond of man pages, even though I can usually dig out what I'm after eventually.
There are a lot of situations where a GUI is preferable, but a powerful command line is a great tool to have at your disposal.
Another example: For a personal hobby project, I needed to make some tools to help me figure out how some text was encoded. I wrote some command-line tools using
I ended up doing a quick and dirty solution in Excel (quick and dirty being relative since I had to implement binary XOR in VBA =P), but if this were something I'd be using frequently I'd make a proper GUI app out of it.
"...always new atoms but always doing the same dance, remembering what the dance was yesterday." -Richard Feynman
Monad turns the command line into an object oriented environment where instead of having to do error prone parsing through text piped though app after app, you treat the output from one app as one or more .NET objects on which you can execute methods, examine properties, and pass them to other applications for further processing.
This is, in fact, far ahead of anything currently available on Unix or Windows.
You mean like Perl? People treating a OO language like the second-coming of Christ. Geesh, shit worked without being totally OO. Perl is great language and it has been doing what you just described since 1987 which is far earlier than .NET