On 19/02/14 19:11, Gian Uberto Lauri wrote:
> Scott Ferguson writes:
>  > On 18/02/14 22:17, Gian Uberto Lauri wrote:
>  > > Scott Ferguson writes:
>  > >  > 
> http://lists.freedesktop.org/archives/wayland-devel/2010-November/000097.html
>  > > 
>  > > I have to check wheter the "seamless" X11 support in Snow Leopard is 
> still
>  > > not so "seamless" as it was Panther (i.e. character rendering, sometimes
>  > > close to unreadable). Gotta check this evening.

I'll leave that to you - I don't pay a lot of attention to Apple, or
Windoof, as they're not particularly relevant in my sector.

>  > 
>  > I'm not sure Apple is an appropriate reference for Linux, but then also
>  > you've previously made the point the Linux is not POSIX and seem to find
>  > that a problem (GNU?).
> 
> Are you sure I did that statement?

No, I wasn't digitally signed. But it did have your distinctive
X-echelon-food: header. :)

> 
>  > I don't see why Linux shouldn't lead instead of
>  > follow flawed models - but it's not my call.
> 
> I agree with you. The problem for me is that it seems to follow these
> models.
> 
> I got your quoting as some form of endorsment of the document at the
> first link. Re-reading your message there is nothing confirming and
> nothing denying.

It (my referencing it) was unclear :/  (life doesn't permit me the time
I'd like to spend answering list posts - or that they deserve).
The whole thread is most enlightening as it encapsulates issues
(Wayland/network).


> 
> Anyway: the reference  to Apple was due a statement  in that document,
> and I did a  test this morning, and X11 support  in Snow Leopard... Is
> the same as in Panther.
> 
>  > > I  am  not  sure  I  would  use  such  an  architecture  for  a  local
>  > > application, at least today.
>  > > 
>  > > That would mean to use some sort of application server, the
>  > > application within the application server and then a browser to run
>  > > the app.
>  > > 
>  > > Very heavy for the machine. 
>  > 
>  > Not necessarily. e.g. https://apps.rutgers.edu/novnc/
>  > http://www.cybelesoft.com/thinrdp/
> 
> It's a VNC that uses a browser...  Nice, but VNC like stuff is not a
> replacement for X. 

X serves both local and remote needs - and it's difficult to separate
the two. That makes development difficult. While the Multix/Unix (I know
the joke) endeavored to satisfy all future needs there are some it
didn't anticipate (micro/mobile devices and their power limitations).

I only quickly referenced a few alternatives - WebRTC is something I
find very interesting. If any alternatives fully replicated X they'd
probably include it's failings. Being able to natively render remote
applications and/or environments (not merely display) is an advantage,
not a failing (IMO).

> I would like to see how much power is required to
> run it.

Runs fine across Australian 3G "broadband" (two tin cans a bit of string
and the attorney general playing man-in-the-middle) with an old Thinkpad
T22 (256MB RAM) as the client. Server requirements are minimal, and can
use a web server for browser clients. Exact figures would depend on
server screen size and plasma plugins.

> 
>  > On the few occasions that I need more than a remote terminal I find x2go
>  > has more useful features and uses less resources than x forwarding and
> 
> Less resources... Which resources? Bandwidth? Memory?

Both.

> 
> X11 is weak when you start displaying large bitmap images (as weak as
> the underlying network link). I did run the first set of OpenMoko apps
> doing remote display through a ethernet-over-USB link and they worked
> fine (it was the telephony that was junk).

Latency is a major problem with bitmap over network, and bitmap is a
very inefficient graphic to begin with.

> 
>  > > I know that HTML 5 can do wonderful things. I am working on a program
>  > > thad does HEAVY use of html 5 and javascript, a program meant to run
>  > > on either the desktop or a tablet. I had to beef mine to 8G mostly
>  > > because of the JavaScript/HTML5 part of the architecture
>  > 
>  > I'm definitely not an expert on javascript[*1] - but 8GiB RAM
>  > requirement seems like it could use some optimising. Have you made use
>  > of HTML5s full local capabilities?
> 
> The problem was the browser  that grown its memory footprint, possibly
> due  the changes  in the  code. Google  guys are  known for  very high
> javascript skills, but belive me, Google mail is not half as complex as
> the application I am working on. It's a simple mail client, not part
> of an ERP system for the health care system :).

The only ERPs I'm familiar with have significant room for remote client
improvement - the two main commercial offerings and the three Open
Source ones (though OpenERP is barely Open Source).

>    
>  > > Then what? two output modules for the same program? 
>  > 
>  > Huh?
> 
> If you do not use HTML5 for both local and remote output then you need
> two output modules.

Only if you chose *not* to use HTML5 to render the remote output. I'm
struggling to think of a valid reason for that - for what it's worth.
Even non-gui clients can scrape the code.

> 
>  > > By  the way,  the application  I  am working  on relies  on a  certain
>  > > implementation (webkit) and does not run on, say, firefox.
>  > 
>  > Then you don't ask anyone else why some applications don't respect
>  > standards. :)
> 
> Sadly I am not in charge neither in the architecture nor in the licensing...

That, in my experience, is exactly the problem (sigh). Blame the cook
for following the owners directions?

> 
>  > > Wonderful, HTML 5  succeeded in turning the clock about  20 years back
>  > > when IE3  understood (non  standard) tags that  Netscape did  not (and
>  > > vice versa).
>  > 
>  > I'm not sure what you're trying to say there... HTML4 is a standard, but
>  > not all browsers support all the tags recognised by the various browsers.
>  > HTML5 has only recently (4th Feb) become a W3C *Candidate*
> 
> That is. We are back to the mid '90. How long will we need to wait until
> an application will run on mozilla or webkit implementation w/o problems?

Until the owners realise how much the lose because their chosen
implementations exclude potential business or lock them into vendors who
ultimately control their profitability? Until hell freezes over? By
profit I mean efficient use of resources - so it applies to charities,
non-profits and "home users".

> 
> That's o.t. with kde and kdm. Thanks for adding the tag :)


The OT tag came with the reposting of the original, um, enquiry on the
debian-kde lists.

Thanks for the though provoking questions.

Kind regards



-- 
To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org 
with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org
Archive: http://lists.debian.org/53046fef.4030...@gmail.com

Reply via email to