Originally posted by: Nothinman
Rosenberg speculates that that might be because "e-mail truly is the killer app regardless of platform," or that desktop Linux still needs a quality e-mail application.
For the most part Evolution and Thunderbird are good enough replacements for Windows MUAs, but it's largely irrelevant though because most places use something like Exchange or Groupwise and thus will use the client that goes with that. With Exchange they could use OWA and forget about the MUA, but I would imagine not everyone would like using a web browser for that.
ya.. It seems like the main barrier for Linux email stuff is that people are using Exchange right now.
Exchange sucks and it doesn't integrate with anything that isn't Microsoft. So if you want to use Exchange then Linux desktop is a pain the ass.
Luckily for me they've avoided using Exchange and use a all linux setup for email. (which is very good at spam control and anti-virus btw, depending on how you set it up)
In that situation they've discovered here is that Thunderbird is easier for people to use then Outlook or Eudora (which is what they used before thunderbird came along). Contact lists and such is taken care of thru a ldap server. Still there is no calendering or anything like that yet...
Seems to me that calendering, contacts, and email is what people use 90% of the time when dealing with Exchange.
Originally posted by: bersl2
Newsforge has a better article IMO. It also points to an infinitely more exciting matter---exciting, due to the fact that it is a flamefest---of
binary kernel drivers. Also, if you then look around Greg KH's weblog, you see the story behind the story, which is
the aloof relationship between OSDL and some kernel hackers.
Official binary driver support for Linux by the Linux kernel developers would be a complete disaster. Horrible horrible thing.
Ya, you could compile ndis drivers and have that work a couple months from now when you do a kernel upgrade... but that would be about it.
If Linux was to support closed source drivers what would happen is that hardware developers wouldn't release binary drivers anymore then they already are...
What would happen is that they would release NDAs to a handfull of developers. Those developers would then turn around and form a company and charge people for those binary drivers to support development of said driver.
This is what happenned in the past for Unix workstations and it would happen for linux.
You'd have to pay 30 bucks to get your sound card working. You'd have to pay 50 bucks to get your video card working. You'd have to pay 15 bucks to get your wifi card working.
This is because Linux would still be a 2% minority OS and these prices would ofset the cost of development and support.
None of them would work out of the box. All of them would require extra configuration and you'd have to make special allowances for them when you do security updates for your system. (basicly would resolve itself is that you Don't do security updates for the kernel) It would be a huge PITA.
Then when you would download that ATI driver from ATI's website it would still crash your computer and be a huge pain, because it's not the Kernel Dev's fault that ATI can't make good linux drivers.
Hell the Linux folks would write the ATI drivers FOR THEM if they would just release the correct documentation for hardware operations.
If that doesn't bother you, then that's fine. Solaris is hoping it doesn't bother lots of people because that is the direction things are going to go with OpenSolaris and those things WILL be supported.
Driver support is not a barrier for business desktop.
You want a video card that works in Linux? Done.
You want a wifi card that works in Linux? There is several and they are cheap.
You want a sound card that works in Linux? There plenty.
SCSI raid? done.
so on and so forth.
Not ALL hardware works out of the box in Linux. But you can find everything you need and it's effectively a non-issue.
For home desktop it's a little bit more difficult. But there are plenty of other more pressing issues.
