Algebra Rizing
in [Lisp]
|
From: Kenneth Tilton on 2 Jul 2010 09:13 Alessio Stalla wrote: > On Jul 2, 11:33 am, Kenneth Tilton <kentil...(a)gmail.com> wrote: >> Andrew Poulos wrote: >>> On 2/07/2010 6:28 PM, Kenneth Tilton wrote: >>>> Asen Bozhilov wrote: >>>>> His kennyness wrote: >>>>>> You do not even know what my math editor does and are smart enough to >>>>>> talk about it anyway? Will you be my friend? i could learn so >>>>>> much....PWUAAHHHAHAHAHAHAHAHHHAA!!!!!! >>>>> I always hate stupid people like you. >>>> It's good to have a steady policy on important issues. >>>>> cljs is strange place. Here come >>>>> different types of gurus ... >>>> What I see, in all honesty, is a sad place where a few people with an >>>> axe to grind over popular JS libraries sit around bullying anyone who >>>> tries to discuss (or learn about or get help with) those projects. >>> I didn't actually believe the negative remarks about the popular JS >>> libraries until I had the temerity to use one in an elearning course I >>> was hired to build. >>> "What could the harm be, its not like I'm controlling a nuclear >>> reactor?", I thought. >>> Alas, while the actual building was fairly painless the reports from the >>> field started coming in about screen freezing, about data being lost, >>> about slowness and delays... None of which occurred during my testing >>> with the "major" browsers on the more common OS. >>> In fact I couldn't reproduce most of the issues. In the end I had to rip >>> out the library code and write my own and so now everything runs fine. >> Right, which is why no one is using JS libraries but me. <sigh> > > Which popular web applications use qooxdoo? Or smartclient? > > Imho, JS libraries for "RIAs" are complete bullshit. Is that why they work so well? Cue the Markovian Hounds... > They basically > rewrite everything down to the layout manager, not leveraging the > browser at all... and in a language, JavaScript, which is interpreted > and single-threaded! And JIT compiled these days!! Try solving -3x-2>13 for x over here: http://teamalgebra.com Now tell me, was the application in fact too slow for you to do Algebra assuming you were learning it? Or are you just arguing from general principles? The latter is dangerous because it means burning hours to speed things up which might not need it -- programmers guess badly at what will be slow, and usually the problem is in the higher-order design, not the implementation stack. If it was too slow for you, go here and do a speed test to Washington, DC: I get 12.43/2.10 down/up. > If you only use the browser as a canvas to paint > on, you'd get much better results with a Java applet. Or Flash - I'm > biased towards Java because that's what I know, My heart goes out to you. > but a few competing > technologies are out there for you to choose. > JS is good for lightweight scripting, not for implementing GUI > libraries. Ah, you /are/ working from general rules. Tsk tsk! Your problem is that I have used qooxdoo enough to see how well it works cross-platform and how fast I am porting a desktop application to the web. ie, You are trying with your kind recommendation to solve a problem that does not exist. kt -- http://www.stuckonalgebra.com "The best Algebra tutorial program I have seen... in a class by itself." Macworld
From: Kenneth Tilton on 2 Jul 2010 09:20 Hang on... Gildas wrote: > It's quite annoying to see dozen of "blank" requests when I > try to write a number... > "dozen[s] of"?! (guessing at the idomatic expression you meant, but even one dozen is a problem because...): There is one xhr per digit. I doubt you were writing a number 24 digits long. Please explain your misreport. Be careful: this will be a commercial product so there is liability for damages involved.* kt * Nah, I just killfile people making silly comments so I can concentrate on serious folks. k -- http://www.stuckonalgebra.com "The best Algebra tutorial program I have seen... in a class by itself." Macworld
From: Gildas on 2 Jul 2010 10:13 > Gildas wrote: > > It's quite annoying to see dozen of "blank" requests when I > > try to write a number... > > "dozen[s] of"?! (guessing at the idomatic expression you meant, but even > one dozen is a problem because...): There is one xhr per digit. I doubt > you were writing a number 24 digits long. Sorry for my bad English, I'm trying to do my best to improve it. > Please explain your misreport. Be careful: this will be a commercial > product so there is liability for damages involved.* Use case : Try to type a digit without numeric keyboard and without CAPS LOCK on. Then, you'll need to press the shift key. Let's say you press shift key one second, you have sent 20 XHR if keyrepeat is set to 20 on your OS.
From: Kenneth Tilton on 2 Jul 2010 10:35 Gildas wrote: >> Gildas wrote: >>> It's quite annoying to see dozen of "blank" requests when I >>> try to write a number... >> "dozen[s] of"?! (guessing at the idomatic expression you meant, but even >> one dozen is a problem because...): There is one xhr per digit. I doubt >> you were writing a number 24 digits long. > > Sorry for my bad English, I'm trying to do my best to improve it. Your english is fine, I just did not want to assume out of hand you meant "dozens" when you had typed "dozen". > > >> Please explain your misreport. Be careful: this will be a commercial >> product so there is liability for damages involved.* > > Use case : > Try to type a digit without numeric keyboard and without CAPS LOCK on. I reach up to the fourth row of my keyboard and type the digit without using any modifier keys. Are you using a cell phone to do Algebra? That is so cool! > Then, you'll need to press the shift key. Let's say you press shift > key one second, you have sent 20 XHR if keyrepeat is set to 20 on your > OS. Oh, that. If you turn off your XHR viewer do you notice a problem? Does your ISP charge you per XHR? Why are you watching XHRs? Are you just looking for a reason to get annoyed? I can prolly avoid that by watching for key-up and key-down events and then keeping track internally (which is how my app works when talking to tcl/tk). Hmm, maybe I can Just Filter Them on the client--I think the modifiers state comes along with the actual key event. Note that this is not an exercise in making comp.lang.javascript happy, this is an exercise in improving math proficiency by moving a little-known desktop application for Windows to the Web. It seems quite fast to me, and the only people who say it is slow also turn out to be library haters whose reports cannot be reproduced. Plonk. If it makes you feel better, performance is indeed a mission-critical concern in my mind and I have my eye out for problems. An initial big concern was indeed the per-key round-trip. So far I have seen zero problems. Your pain is self-inflicted: you are watching XHRs for no reason. kt -- http://www.stuckonalgebra.com "The best Algebra tutorial program I have seen... in a class by itself." Macworld
From: Alessio Stalla on 2 Jul 2010 10:51
On Jul 2, 3:13 pm, Kenneth Tilton <kentil...(a)gmail.com> wrote: > > Which popular web applications use qooxdoo? Or smartclient? I'm waiting for an answer... > > Imho, JS libraries for "RIAs" are complete bullshit. > > Is that why they work so well? Cue the Markovian Hounds... Evidently I don't believe they work that well... > > They basically > > rewrite everything down to the layout manager, not leveraging the > > browser at all... and in a language, JavaScript, which is interpreted > > and single-threaded! > > And JIT compiled these days!! Try solving -3x-2>13 for x over here: > > http://teamalgebra.com > > Now tell me, was the application in fact too slow for you to do Algebra > assuming you were learning it? Or are you just arguing from general > principles? The latter is dangerous because it means burning hours to > speed things up which might not need it -- programmers guess badly at > what will be slow, and usually the problem is in the higher-order > design, not the implementation stack. I'm not talking specifically about Algebra, and I'm not talking only about speed either, but about general user experience. > > If it was too slow for you, go here and do a speed test to Washington, > DC: I get 12.43/2.10 down/up. > > > If you only use the browser as a canvas to paint > > on, you'd get much better results with a Java applet. Or Flash - I'm > > biased towards Java because that's what I know, > > My heart goes out to you. I'm not saying Java is perfect, far from it, just - in my humble opinion - better than Flash and Silverlight. We all know that if the browser was written in Lisp we would have far less problems :) > > but a few competing > > technologies are out there for you to choose. > > JS is good for lightweight scripting, not for implementing GUI > > libraries. > > Ah, you /are/ working from general rules. Tsk tsk! > > Your problem is that I have used qooxdoo enough to see how well it works > cross-platform and how fast I am porting a desktop application to the > web. ie, You are trying with your kind recommendation to solve a problem > that does not exist. The problem is precisely "porting a desktop application to the web". That cannot be done, period. Not without either changing the application to embrace the "web way", or using the browser as a mere deployment tool, not a platform. Heavy use of AJAX + redoing a whole GUI library in JS just doesn't cut it, at least today. Maybe in a few years browsers will evolve, new protocols will be developed, and it will indeed be possible to have desktop-like applications on the web, but not today. Alessio |