Sorry guys Not sure I'd fully agree with all your set forth reasons. I think it's more of a distinction between the Worse-is-Better approach to programming in the (mostly) M$ world, as compared to the "The Right Thing" approach in most of the Linux software. It's also a "feature" of the different approaches between commercial (WiB) and open source software (sometimes TRT).

The idea behind the WiB approach is to make your program just about usable, then sell it on to Joe public. You can later apply fixes and upgrades to cover the shortfalls due to the rushed sales date. The MIT approach tries to make the program as complete and bug free as possible before presenting it to the user. The reason most commercial programs follow the former approach is there's more money in it: While your product is inferior, you're at least obtaining income, and due to the upgrades you keep on getting income as you improve the program over time. If they followed the MIT approach, they'd have to wait until they've finally completed the near perfect program before even thinking about income - which could be decades without any income, while your competitors are already reaping huge profits.

As for the multi-tasking, sharing, etc. I'd be much more willing to agree to those statements.

As for the M$ origins (i.e. DOS) ... it's not as clear cut as that - DOS was based on CP/M ... And from that there are several stuff still in existence in modern Windows: e.g. 8.3 short filename convention, console wildcards, reserved file names like PRN/CON/COM1/etc. That said, the NT branch of Windows (on which all modern windows after Win95/88/ME is based, i.e. WinNT3/4/Win2k/WinXP/Vista/7) actually originated from the IBM OS2 operating system. The reason NT had (and still has) hiccups is due to its attempt at backwards compatibility to the old DOS programs. This has been less in evidence since Vista - you'll find many older programs simply will not run on the new Windows. Even most DOS programs have difficulty running on Win7 (you could use the DosBox emulator for this though). So it seems M$ is going away from the yoke of backwards compatibility, it might be a good thing, but I'm not holding my breath

I'd most likely would want see "gamemaking" software being ported. There is lack of such and if you want to make game, but for some reason can't program (like me - I can code normal applications like notepad, word, even graphic software or database one, but I'm unable to make game, because of my helplessness when it's come to collision handling or "writing" levels in code [could use level editor like tiled or make one, but there's always problem with writing importer/level loader/saver]) or just don't want to you have very limited choice.
There's Stencyl (good for flash games, but using it to wrap game into flash projector and use it stand alone is a little out of purpose, isn't it. Although Stencyl in 2.0 will support native export. We'll see how it'll turn out), Game Develop (MMF/Construct-like, although official Linux download is ubuntu-only, though there is test build for other distros somewhere in the forums) and so-called Game Editor that is failure and simple no-go for both coders and game designers that simply want to make game easy. For coders, because it's built-in level editor sucks and you'll better with writing your own/using tiled and code game from scratch because it doesn't add any advantage that fastens development and for it's internal coding language use Objective C, meh. For game designers, because you cannot click-out game and again, level editor sucks and you'd better with other program like Stencyl or Game Develop.
There's also lot of half-assed "gamemaking" (quotes on purpose) software like Game Baker or ika, that are better to be forgotten.

Long story short, I'd like to see Linux ports of Game Maker, Clickteam's TGF/MMF and Unity. For last it doesn't have to be actual unity IDE, may be just webplayer/runtime so you can play it on Linux.

//edit @irneb: I think console wildcards are good and should be implemented in GNU userland as well. Of course regular expressions are good, but for more complicated things. Sometimes all you need is <knownpart><character/partial string that you don't know><knownpart> and there wildcards are better than regular expr. because it's easier to construct expressions with wildcards (only two special characters: * and ?) than with regular expressions (many special characters which give you headache when you try to remember them). Of course since Linux/Unix allows asterisk and question mark in names files with those characters at specific position should be also taken into account, e.g. when you write ls loveletterfor*, there should be displayed loveletterforanya, loveletterformikka, but also loveletterfor*. And before you ask: Yes, I know ls does support wildcards and does take names with wildcards into account, but many other tools in GNU userland doesn't, or at least doesn't support them fully. E.g. I can't write cp *.txt *.asc to quickly make copy of all txt files, but with asc extension, like I would do in Windows.

But that's, of course is off-topic.

Vsphere client & Server
Photoshop
paint

notepad ++ FTW

notepad? What's wrong with vi?

I think he was referring to the program Notepad++, which is significantly more powerful than notepad (unfortunate choice of name if you ask me). And although I prefer vim or emacs, I can understand why someone unwilling to invest the time to learn these editors might want notepad++. It is probably the most powerful editor that has the standard windows interface (i.e. primarily mouse interaction for commmands). I agree that it is unfortunate that notepad++ isn't available for linux, but there a many, many powerful editors available for linux, and if you do some heavy coding, it might be worth learning emacs or vi(m). And geany and scite are similar to notepad++ I believe.

Yep, I'd also recommend SciTE. It's lighter than N++, but has most of the same features. The addons community, isn't as large though, so not as many extra addons built specifically for ST. But it is based on the Scintilla editor library, which does have a great following.

Even on winsnooze, this is my editor of choice. Especially for source coding in lisp.

+1 vote for photoshop. It's a must-have.

OpenBSD Packet Filter

I've been a user of Wordperfect for close to 20 years.
It is the only wordprocessor I use.
Would be great if they would reopen WP6, which had a linux version.

Kate & Kwrite = notepad. Open or Libre office = Word Perfect. + kPDF writes PDF docs for free. GIMP = Photoshop for most pics.

Perhaps, though being a Notepad clone would not be something to write home about!

Definitely in disagreement with this. Perhaps you meant: "Open or Libre office = Microsoft Office". Word Perfect is literally years (if not decades) in advance of OOo.Write / MS Word.

As does several packages based on the open source GhostScript tools, some create a virtual printer-to-pdf so any of your programs can simply print to PDF instead of SaveAs to PDF. Even OOo has a built-in PDF save-as feature, as do Word Perfect, even as do MSO 2007/2010.

Up to a point. I'd say you're correct by inserting the "most" qualifier. CMYK is at its infancy in Gimp, but then Gimp has other features which PS is deprived of. So it depends on your "image" I suppose.

Even OOo export to PDF and print PDF via CUPS work fine in my experience - but I'm not making super complicated, charted documents. I should try Libre office one of these days.

As for notepad, I've been happy as a clam with gedit and Bluefish (I know comparing Bluefish is pushing it a bit, but for plain text / non WYSIWYG web work, it's quite nice), so thought I'd put that combo out there in case they haven't been mentioned together in the last 4,000+ posts.

Do it! There's no reason to stay with OOo. LO looks and feels exactly the same - so no learning curve ... only it's much more stable than OOo. So AFAICT it's like a final release (LO) compared to a beta (OOo). There's not a lot of extras in LO over OOo, here-n-there you'd come across a feature (built-in) which used to be an addon in OOo, but even this is few and far in between.

Good news on Serv-U secure file transfer server. RhinoSoft ported Serv-U to Linux in version 11 (http://www.serv-u.com/releasenotes/) and then joined Red Hat's ISV program (http://www.rhinosoft.com/press.asp) to ensure it worked on RHEL.