Date/Time: Thu, 18 Oct 2018 17:20:05 +0000
Post From: Future planned OS platforms for SC
|bjohnson777 (Brett Johnson) - Posts: 237|
User35525 nicely covers the OS market share charts. I'll point out a few extra things.
Winxp still being so high in the list tends to point to a corporate usage that is unwilling to go to new win versions... probably for reasons already mentioned. That is not a positive for M$, especially after winxp was EOL'd and not getting any security updates. There seems to be a high resistance to win8 and win10 in the corporate world.
In big corporations, the average user/employee has no say about what computer and OS is being used. M$ still has a monopoly in that market and will skew the statistics.
Ganz using video game and PornHub articles in an argument is a kinda desperate.
Andreas reinforces the common argument of M$ quality concerns.
Making a linux (or other OS) version of SC shouldn't really be looked at as porting and maintaining a separate branch. It should be looked at as a global upgrade that adds multiple platforms as part of the upgrade.
Getting an SC version on one of the open unix platforms would be a major boost to reliability and perception. Many of the high dollar traders won't put up with M$ defecation and rot because it isn't cost effective for so much time to be wasted. Those traders may have the admin skills required to set up an open unix on their own. If not, they have the money to pay someone and take it as a business expense.
One reason I keep mentioning the BSD's is that they are used by pretty hard core unix admins. The BSD's ship much lighter weight than the consumer linux variants. They tend to be more secure than linux. Since Mac is a BSD variant, getting SC on BSD shouldn't be much extra work. Eventually I envision hosted headless versions of SC trading robots similar to what MetaQuotes does with their ExpertAdvisors. Having a BSD version of that clearly states: "We don't mess around. If you want a serious and professional option, it's right here."
Since QT gets bashed the most, let's use it in an example analysis. Yes, there are some reported bugs. The main questions should be:
* Are these bugs in the main core or peripherals?
* Are these bugs being actively fixed?
* What is the expected fix time?
* Are the peripheral bugs relevant to a charting package?
* Are there simple workarounds for the bugs related to a charting package?
* Since QT is open sourced, is it worth the SC dev's time to fix some of the bugs and submit patchs?
* Given all this, are the very wide cross platform gains enough to offset the time working around the bugs?
These questions then get asked of WX and the others User35525 listed. These questions can also be asked against the various win versions (excluding the last two).
"Xwindows system is outdated and is not fully controlled by devs"Your pro M$ argument about this statement hardly holds water. This is just you whining again about not being able to play every video game under the sun on X11. The X11 system in general was designed for networking and multi-user environments on mainframes. While it has some quirks, it is hardly outdated. The general accelerated driver interfaces are more than enough for a charting package. The real bottleneck in a charting package isn't drawing lines on the screen, it's doing the calculations for all those lines in an efficient manner. Some indicators have very heavy calculations per bar. It's something I run up against constantly in my own tool kit DLL and my own scratch pad linux program.
"Xwindows system is outdated and is not fully controlled by devs"News flash: THE SC DEV'S DO NOT CONTROL M$ OR APPLE! If there is a problem, they go to the back of the line like everyone else. Linux and BSD X11 is open sourced. If the SC dev's really wanted to, they could fork their own version and control it to their heart's delight. I don't see the need for this, though.
A quick add to something I said previously: Releasing your own commercial versions of win and osx without absurd licenses can get you thrown in jail. Nobody cares about releasing your own version/flavor/tweaks of the open unix's. A custom stripped down linux live CD that could also be used as an installer disc could have basic security settings and performance tweaks already in place. This alleviates the need for the end user to be admin level saavy.