How much of one's work day can be handled by browser based software. Can we get it to 100% ?
I wonder if the path of moving everything to a server except for a local terminal browser is practical? As in having all the applications be web browser accessible which makes geographical location functionally irrelevant. All that matters is a net link from the Linux based browser to the server/s. Most of the hassles of local user support go away. Yes- we could argue about Dumb Terminals reincarnated or perhaps just examine the server concepts. The reason for my mentioning this comes from an observation on hardware already out in the wild. We have a deluge of computers being replaced largely because of gaming and bloatware. Arguing about the elements of a distro suited to lightly load hardware have been beaten beyond dead. The concept of what older hardware could serve as servers- pun potential eh? is the new question. I see it as we take that 2ghz 2gig ram "last year's model" and stuff several 500 gb drives in it for a family or small office server. Then the users take whatever suits their fancy for a terminal-from the 7K alienware gaming laptop down to a $5 from surplus exchange T30 thinterm. Thus whatever disaster befalls their "local hardware" the data is safely replicated on their server's drives and perhaps mirrored elsewhere. For a reason. It's how we can make the typical users have a bit more assurance that their sweated over data is safer than other ways we've been doing this. As reassuring users that their "world" is safer being distributed so to speak gets easier if it's true... The Long term path I see is colo's having a farm of replicant application web servers to mirror local servers.. With enough hot replication that the user's sessions etc are "safer" for it.
On Mon, Nov 3, 2008 at 4:11 PM, Oren Beck orenbeck@gmail.com wrote:
I wonder if the path of moving everything to a server except for a local terminal browser is practical? As in having all the applications be web browser accessible which makes geographical location functionally irrelevant. All that matters is a net link from the Linux based browser to the server/s. Most of the hassles of local user support go away. Yes- we could argue about Dumb Terminals reincarnated or perhaps just examine the server concepts. The reason for my mentioning this comes from an observation on hardware already out in the wild. We have a deluge of computers being replaced largely because of gaming and bloatware. Arguing about the elements of a distro suited to lightly load hardware have been beaten beyond dead. The concept of what older hardware could serve as servers- pun potential eh? is the new question. I see it as we take that 2ghz 2gig ram "last year's model" and stuff several 500 gb drives in it for a family or small office server. Then the users take whatever suits their fancy for a terminal-from the 7K alienware gaming laptop down to a $5 from surplus exchange T30 thinterm. Thus whatever disaster befalls their "local hardware" the data is safely replicated on their server's drives and perhaps mirrored elsewhere. For a reason. It's how we can make the typical users have a bit more assurance that their sweated over data is safer than other ways we've been doing this. As reassuring users that their "world" is safer being distributed so to speak gets easier if it's true... The Long term path I see is colo's having a farm of replicant application web servers to mirror local servers.. With enough hot replication that the user's sessions etc are "safer" for it.
I hate this move to web based everything. It's a supposedly cheap way to build software, where in fact way too much time is wasted on UIs. There are way to many ways to build cross-platform client-server software to have everything be web based.
-
Arthur Pemberton -
Oren Beck