All being said about X Bloat With Some Denial Thereof, I should perhaps make my "feelings" about X a little more clear.
X has various flaws
The fact that it doesn't provide any "hints" as to what the user interface should look like has proved unhelpful.
The lack of GUI policy has resulted in an unfortunate proliferation of sets of GUIs and GUI policies.
This establishes a bit of a dilemma: It would be preferable if there were a dominant GUI toolkit, but none has been sufficiently superior to the others as to become dominant.
Authorizations are extremely painful to configure.
X's font handling isn't great. X doesn't know about antialiasing, which is a concept that postdates its design.
The X system design is over ten years old and X could do with a substantial redesign based on the learning that has taken place from the whole process of building X as well as various other graphical environments.
Unfortunately, the redesign and redevelopment process would not take substantially less time than the initial design and development, and I would suggest that it would represent a project involving on the order of hundreds of millions of dollars worth of effort.
Despite the flaws, I find X useful
There are a lot of useful applications that have used its nearly infinite malleability. Running apps across networks is really useful. The fact that it simultaneously accommodates applications using different GUIs is useful.
I haven't yet seen anything that reasonably replaces X
Most typical is the scenario where someone proposes an alternative that wouldn't represent an alternative unless the sorts of people presently working on X leaped over to the GnotX alternative, and did a couple of years worth of development.
Various would-be replacements have promoted themselves; while they may have provided improvements in one area or another, none have provided all of the following:
Support for legacy applications that presently use X
Support of remote display of applications
I don't want a replacement that simply throws away all the existing code.
I would be perfectly happy to migrate to the "latest and greatest" windowing system, so long as I can continue to use EXMH , GIMP , cbb , Netscape , and ApplixWare . And hopefully WINE .
Proponents of alternative systems tend to assume some combination of the following:
Everyone will be happy to throw away all the existing X applications, or
Applications will only run on Some Particular Platform.
Some vendor will support all sorts of OS and hardware platforms, in perpetuity, and be perfectly trustworthy to provide economical licensing terms to all.
There have been no alternatives to X that provided suitable licensing to allow them to be universally adopted.
People denigrate X based on one feature or another being inferior to those found in other systems. This is only really useful if it is possible to adopt that other system, or if the other system does not have global flaws that make up for some local advantage.
MacOS may have some nice "Interface Design." But since Apple has tightly restricted its licensing, it does not represent a universally usable solution.
Similarly, Apple controls NeXTStep . It may become an interesting alternative, if and when GNUStep becomes mature enough for deployment of significant applications.
Many things have been hacked onto Win32; unfortunately, actually using Windows-based facilities has two massive detrimental effects:
Since Windows is a giant hack that more resembles a giant ball of duct tape built on top of CP/M than anything else, reliability suffers incredibly, and efforts have to work around the hacks.
You have to trust Microsoft.
There are a number of things that I would trust Microsoft with. I trust Microsoft to continue to release buggy software such as the Windows 98 release that needed patches to survive to 1999, let alone 2000. I trust Microsoft to continue to release software that gets more bloated such that it requires as much memory as systems can support just to support useless chrome.
But control over whether my code will continue to be usable next year whether due to technical or licensing changes is not one of the things I'd like to trust them with.
There are various "niche" systems such as Be, QNX / Photon , MGR , GEOS, NeWS, and CLIM. They may have had valuable concepts that have been lost due to nonadoption. The community as a whole can't adopt them if they are only available on "niche" platforms, or under highly restrictive licensing constraints.
I am not offended or shocked by the notion that someone might create something better than X. I think it unlikely that the development of an alternative would require substantially less than the hundreds (perhaps thousands) of person-years that were expended in making X as functional as it is today.
I am offended by the notion that everyone should, as a result of this better system, stop, and throw away all old applications and rewrite them for the new system, especially before it has been clearly established that the new system is superior to X across all areas of functionality. X satisfies needs that no other system can presently satisfy.