Algebra Rizing

[Kenneth Tilton]

Kenneth Tilton schreef:

You credit yourself too much.
Even though you behaved like a horny baboon in here, I seriously doubt
you triggered anybody in here enough to explain the DDOS attack.
(As if we all have zombienets at our command...)

You mostly got yourself plonked I think...

Erwin Moller

I suppose. I happened to be watching and saw all these key events coming
in on a math widget and they were all "x" so I figured someone was just
tryign to break the app by holding down the x key, and then it turned
into a stream of a hundred x's at a time. It looked specific to the app,
but what do I know? Maybe that is how bots work, look for something that
responds and then pound it. It does sound like fun.

Not a big concern.

kt

 

[Kenneth Tilton]

You credit yourself too much.

[...]
I suppose. I happened to be watching and saw all these key events coming
in on a math widget and they were all "x" so I figured someone was just
tryign to break the app by holding down the x key, and then it turned
into a stream of a hundred x's at a time. It looked specific to the app,
but what do I know? Maybe that is how bots work, look for something that
responds and then pound it. It does sound like fun.

And that after you've been told a number of times that sending HTTP
requests for every single key event is a bad idea. Your application got
itself DoS'ed by a single user who held down the "x" key too long,
possibly even by accident, but you still managed to blame it on the
group. Congratulations.

Your "solution" for the client side was to not look at the requests.
Maybe your server will feel better if you don't watch the logs?

In summary you and the other JS library haters now find yourselves
sitting around rejoicing in an obviously hard-working programmer running
into occasional difficulties. You might want to retrace your steps to
find where you went wrong. It's prolly the hate, tho.

kt

 

[Erwin Moller]

Kenneth Tilton schreef:

Erwin Moller wrote:
You credit yourself too much. [...]
I suppose. I happened to be watching and saw all these key events
coming in on a math widget and they were all "x" so I figured someone
was just tryign to break the app by holding down the x key, and then
it turned into a stream of a hundred x's at a time. It looked
specific to the app, but what do I know? Maybe that is how bots work,
look for something that responds and then pound it. It does sound
like fun.

And that after you've been told a number of times that sending HTTP
requests for every single key event is a bad idea. Your application got
itself DoS'ed by a single user who held down the "x" key too long,
possibly even by accident, but you still managed to blame it on the
group. Congratulations.

Your "solution" for the client side was to not look at the requests.
Maybe your server will feel better if you don't watch the logs?

In summary you and the other JS library haters now find yourselves
sitting around rejoicing in an obviously hard-working programmer running
into occasional difficulties. You might want to retrace your steps to
find where you went wrong. It's prolly the hate, tho.

kt

Kenny,

You don't listen/read very well, and thus you won't learn.
That is really your loss, Kenny.

People in here warned you that sending a xmlhttprequest on each
keystroke is a very bad idea (latency issues, but also firing up a
process on the server on each keystroke).

But you can only respond with 'lib haters!'.

When a police officer stops you when you drive your Ferrari at 150
miles/hour through the city centre, you respond: "You all just hate
Ferrari and are jealous I have one!".
I hope not.
But that is how you act in here.

Most people in here responded to your posting without the intention to
irritate you, they wanted to help you, but all you gave back were
insults and wisecracks.

Last thing: Those 'lib haters' as you call them, didn't it occur to you
they maybe have a valid reason for that?
You are so damned convinced you are right that it is impossible for you
to have an objective opinion about the matter.

Erwin Moller

PS: If you respond like a baboon again I promise I will never bother you
again.


--
"There are two ways of constructing a software design: One way is to
make it so simple that there are obviously no deficiencies, and the
other way is to make it so complicated that there are no obvious
deficiencies. The first method is far more difficult."
-- C.A.R. Hoare

 

[Kenneth Tilton]

Kenneth Tilton schreef:

On 05/07/10 22:35, Kenneth Tilton wrote:
Erwin Moller wrote:
You credit yourself too much.
[...]
I suppose. I happened to be watching and saw all these key events
coming in on a math widget and they were all "x" so I figured
someone was just tryign to break the app by holding down the x key,
and then it turned into a stream of a hundred x's at a time. It
looked specific to the app, but what do I know? Maybe that is how
bots work, look for something that responds and then pound it. It
does sound like fun.

And that after you've been told a number of times that sending HTTP
requests for every single key event is a bad idea. Your application got
itself DoS'ed by a single user who held down the "x" key too long,
possibly even by accident, but you still managed to blame it on the
group. Congratulations.

Your "solution" for the client side was to not look at the requests.
Maybe your server will feel better if you don't watch the logs?

In summary you and the other JS library haters now find yourselves
sitting around rejoicing in an obviously hard-working programmer
running into occasional difficulties. You might want to retrace your
steps to find where you went wrong. It's prolly the hate, tho.

kt

Kenny,

You don't listen/read very well, and thus you won't learn.
That is really your loss, Kenny.

People in here warned you that sending a xmlhttprequest on each
keystroke is a very bad idea (latency issues, but also firing up a
process on the server on each keystroke).

But you can only respond with 'lib haters!'.

When a police officer stops you when you drive your Ferrari at 150
miles/hour through the city centre, you respond: "You all just hate
Ferrari and are jealous I have one!".
I hope not.
But that is how you act in here.

Most people in here responded to your posting without the intention to
irritate you, they wanted to help you, but all you gave back were
insults and wisecracks.

Last thing: Those 'lib haters' as you call them, didn't it occur to you
they maybe have a valid reason for that?
You are so damned convinced you are right that it is impossible for you
to have an objective opinion about the matter.

Erwin Moller

PS: If you respond like a baboon again I promise I will never bother you
again.

What are you talking about? I saw a stream of a hundred or so x's come
in faster than anyone can type. Then there was a pause and and an
equally fast stream of requests came in, each with a 100 x's and a
hundred something-else's. This went on non-stop for the 20s or so it
took to switch screens and kill the server.

And you think that is the same as a person actually doing Algebra typing
one character at a time?

That is what this group warned me would not work?

oooh-oooh-ah-ahh-eeep!!!!

kt

 

[Gregor Kofler]

Am 2010-07-06 12:59, schrieb Kenneth Tilton:

What are you talking about? I saw a stream of a hundred or so x's come
in faster than anyone can type. Then there was a pause and and an
equally fast stream of requests came in, each with a 100 x's and a
hundred something-else's. This went on non-stop for the 20s or so it
took to switch screens and kill the server.

So? This only proves the idiocity of your (or qooxdoo's) approach. One
can execute a DoS attack by simply pressing a key. Cool. Besides: I
suppose neither you nor your "application" is in any way worthwhile for
DoS attacks.

And you think that is the same as a person actually doing Algebra typing
one character at a time?

Perhaps he or she left a book on the keyboard...

That is what this group warned me would not work?

oooh-oooh-ah-ahh-eeep!!!!

Once more - bugger off.

f'up set.

Gregor

 

[Kenneth Tilton]

by the way...

What makes you think the server fires up a (lightweight) process on each
request?

And what makes it necessarily a very bad idea if requests and responses
are very small and the server keeps connections open?

So what do folks here recommend in the way of software that simulates
real-world loads, so we can put some numbers on this. I /am/ curious
about how many users one AWS instance will support.

otoh, until I have real students (vs. antagonistic hairless apes) using
the software) I am not sure how I would specify a typical user's pattern
of interaction.

I know you all /want/ me to fail (witness your mistaken glee over either
a DoS attack or a bug in my code that caused The Flight of the Xes), but
if you do not mind I'll optimize according to actual performance, not
imagined. We programmers tend to guess badly on performance.

kt

 

[Kenneth Tilton]

OK, firebug is out, the Training Center is in, and quite a number of
fixes. I'll go look now at cloning the AWS instance to Europe and Cali
so you furriners can check performance.

The Training Center shows implicitly the skill set of the underlying
engine. It also let's you create a bunch of random problems and work on
them with less typing.

Next step is to provide the option to see a problem solved for you,
though only in the training center.

Mind you, for you yobbos this is just an ongoing demonstration of
qooxdoo /and/ qooxdoo driven completely from a Lisp server application.

Yes, lot's of glitches and UI annoyances remain. Those get addressed if
they get in the way badly enough or if they are trivial enough to fix.

Interesting punch-line: I am still learning qooxdoo and I am still
learning Interweb programming, and look what I have put together in nine
weeks working part-time.

http://teamalgebra.com/

Perhaps he or she left a book on the keyboard...

Yeah, but then came the stream of requests each with a hundred x'es and
some a hundred I forget's.

My *nix screen fu is deficient or I would have been able to go back and
look at those, see if they were coming in as recognizable requests,
which would suggest the traffic was observed and then aped. Sorry, I
know primates are a touchy subject for you.

Once more - bugger off.

Love your sense of humor. Hint.

kt

 

[Kenneth Tilton]

You're just as obnoxious as the library haters you complain about.

Good lord, is David Mark the only one in here with a sense of humor?

If it makes you feel better the site does not work from my own computer
(tho it works from others I have lying around) and I serious broke
focus-handling with the last change I made before uploading. Now there's
a shock.

kt

 

[David Mark]

Good lord, is David Mark the only one in here with a sense of humor?

Speak of the devil...

If it makes you feel better the site does not work from my own computer
(tho it works from others I have lying around) and I serious broke
focus-handling with the last change I made before uploading. Now there's
a shock.

Summary:

* Kenny doesn't know anything about browser scripting, CSS, HTML or
anything else to do with "InterWeb programming" (and likely never
will).

* Kenny blames all of his problems on the people who warned him in
advance that he would have such problems (e.g. "library haters")

* Kenny now sees everybody as "library haters" because nobody can seem
to make his application work.

* Kenny inexplicably thinks his (appropriately labeled) "train wreck"
is an example of how well KooksDo "works".

* Kenny gets angry when bugs are reported and chastises reporters for
missing some imaginary "big picture" that is apparently only visible
to him.

* Kenny thinks that "advanced" libraries like KooksDo and Cujo are
going to replace HTML/CSS/JS because he's seen demos that look like
(very slow and unwieldy) desktop applications.

In short, Kenny is a clueless, fantasizing, disagreeable greenhorn
(not at all uncommon in the industry). Just ignore him and maybe he
will go away.

 

[David Mark]

[...]

* Kenny thinks that "advanced" libraries like KooksDo and Cujo are
going to replace HTML/CSS/JS because he's seen demos that look like
(very slow and unwieldy) desktop applications.

And echoing earlier sentiments, I too find it bizarre that these RIA
frameworks, with the stated purpose of imitating desktop applications,
replace standard OS components (e.g. form controls) with incompetently
implemented blobs of HTML and CSS that "work" (minus a ton of the
built-in and free functionality) only in a handful of the latest
browsers (if that).

Typically these things are styled to look like the latest incarnations
of popular operating systems (e.g. Windows 7). Of course, the built-
in controls look (and work) exactly like the OS versions used by
desktop applications because they *are* the same ones used by desktop
applications. Windows 7 form controls look *exactly* like those found
in desktop applications running on Windows 7. Mac form controls look
just like those found in desktop applications running on a Mac. And
when Windows 8 comes out, the built-in browser controls will look and
work just like the controls found in Windows 8 desktop
applications.

So what could the developers of these frameworks be thinking? It's
the same old (outdated) control-freak notion that they must make
everything look exactly the same in every browser. That and they
can't stand the thought of form controls that "clash" with the rest of
their "themes". The fact that the end-users would be much happier
with form controls that look (and work) like the form controls found
in their desktop applications (the same ones these frameworks are
desperately trying to equal) seems to elude them.

And why would they copy Windows 7? Because it is there. You can bet
they'll be feverishly working on a "theme" to mimic Windows 8 when it
comes out. If only they would stop and think about what they are
doing and realize it is all a waste of time.

 

[Kenneth Tilton]

Speak of the devil...


Summary:

* Kenny doesn't know anything about browser scripting, CSS, HTML or
anything else to do with "InterWeb programming" (and likely never
will).

* Kenny blames all of his problems on the people who warned him in
advance that he would have such problems (e.g. "library haters")

* Kenny now sees everybody as "library haters" because nobody can seem
to make his application work.

* Kenny inexplicably thinks his (appropriately labeled) "train wreck"
is an example of how well KooksDo "works".

* Kenny gets angry when bugs are reported and chastises reporters for
missing some imaginary "big picture" that is apparently only visible
to him.

* Kenny thinks that "advanced" libraries like KooksDo and Cujo are
going to replace HTML/CSS/JS because he's seen demos that look like
(very slow and unwieldy) desktop applications.

In short, Kenny is a clueless, fantasizing, disagreeable greenhorn
(not at all uncommon in the industry). Just ignore him and maybe he
will go away.

Thanks, Dave. I was looking for an excuse to jump back in since there is
no news on the JS front to justify me posting here again. It's just
going so well with qooxdoo that nowadays the server-side functionality
is my only concern. As soon as I reach the next milestone (today, I
hope) I'll spam you all about my Algebra software in the guise of
responding to this.

I owe you one!

kt

 

[Kenneth Tilton]

I don't see much interesting in his posts. :-(

Have you looked at qooxlisp? http://wiki.github.com/kennytilton/qooxlisp/

Cells? http://wiki.github.com/kennytilton/cells/

Didn't think so.

How do /you/ program the web?

kt

 

[Kenneth Tilton]

Speak of the devil...


Summary:

* Kenny doesn't know anything about browser scripting, CSS, HTML or
anything else to do with "InterWeb programming" (and likely never
will).

And I can still build this:

http://teamalgebra.com/

In ten weeks. Working only week-ends. Get it?

If you'll notice, that web site let's kids enter any problem and then
solve it any way they like and is able to check any step they enter and
offer hints on how to proceed, even if the software would have done it
differently. That is not the easiest functionality to implement, so how
about you spend all your time fussing with browser variations in HTML
and CSS and HTML while I do something a tad more interesting?

* Kenny blames all of his problems on the people who warned him in
advance that he would have such problems (e.g. "library haters")

I am having problems with qooxdoo? Maybe you think the focus thing was a
qooxdoo "problem"? Sorry, Charlie, my app does things with the focus,
too. Duhhhhhh. The bug was in my handling.

Cells fans will be interested to know that I got into a dataflow ccyle
where the server told qooxdoo about focus changes when the app wanted to
initialize the focus to, say, the top-left widget in a new form but (of
course) qooxdoo has to tell the app when the user clicks on a widget to
change the focus. Normally I get all hairy with code trying to detect
cycles, but this time it occurred to me to do something new in my
experience with external GUIs: the Lisp application, when it wants to
focus on X, does so by having the client eval /qooxdoo/ code to focus on
X. The lisp app then learns about the focus change in the same way as if
the user had clicked on X! Roundabout? Guilty, but possibly Deeply
Correct. We'll see if it holds up.

* Kenny now sees everybody as "library haters" because nobody can seem
to make his application work.

No, I read what they write and I can tell where they are coming from.
Especially when they end by suggesting my editor does not work because I
am using a library. See straw. Grasp.

* Kenny inexplicably thinks his (appropriately labeled) "train wreck"
is an example of how well KooksDo "works".

Only to the intellectually honest observer. That I cannot help you with.

* Kenny gets angry when bugs are reported and chastises reporters for
missing some imaginary "big picture" that is apparently only visible
to him.

It was all a misunderstanding. I hid too well the warning that the site
was offered "as is". That is now the "front page".

* Kenny thinks that "advanced" libraries like KooksDo and Cujo are
going to replace HTML/CSS/JS because he's seen demos that look like
(very slow and unwieldy) desktop applications.

Right, Kenny is the only one using JS libraries. And you call me
"fantasizing"?

In short, Kenny is a clueless, fantasizing, disagreeable greenhorn
(not at all uncommon in the industry). Just ignore him and maybe he
will go away.

Again, thx so much for bringing me back with this post days after the
last in the thread. It would have been really lame for me to try to
resurrect it myself. If you are in the city I'll buy you a drink.

The enhancements in this relase are:

- I re-ran the Algebra engine regression test and found quite a few
things to fix (and fixed them).

- In the Practice Room you can now click "Show Me" and the program will
solve a generated problem step by step. Teachers do not have to write on
the board any more, or turn their backs on the little devils. Click on
any step and the program will explain what it did.

- Miscellaneous other bug fixes.

Non-enhancements:

- No AWS instance outside the East Coast yet. Turns out one can easily
clone an instance but only within a region. I guess my software is not
the only work in progress. The next version might be presentable enough
to take to educators, then I'll worry about covering other regions.

Known flaws:

- Many UI annoyances and confusions

- Hints and explanations refer to actual terms from the problem at hand,
even when entered by the student. But! That is only partly implemented
(it is a tad tricky, that code, and I punted a bit).

Known bizarritude:

- At some point the software will start positioning math too high within
a widget. You'll know if it happens. I have not sorted out what
triggers this, but once it starts it stays that way so once I notice (or
someone complains) I will dynamically set a flag that tells the code to
compensate. Don't you wish you were a Lisp programmer?

Next up:

- the so-called "Mission" module that offers no hints and no second
chances and requires 80% to get certified in a skill.

- "Review" lessons that offer a mix of problems from all the subsections
of a chapter. That code is there, just needs resurrection.

Cue the hounds, who have left as an exercise explaining why it bothers
them so much that a fellow programmer is having fun quickly bring a
ground-breaking application to the Web.

Again, for your convenience: http://teamalgebra.com/

kt

ps. In case anyone does not find this interesting, go to the site and
poke around and meditate on this: I write Lisp all day, and lately
nothing but Lisp. I rarely look at qooxdoo doc (tho that will fire up
again if/when I need a new widget). I almost never look at HTML and CSS
references. Yet I am building a Web application. Ommmm.... kzo

 

[Krzysztof Drewniak]

[snip other things]

And I can still build this:

http://teamalgebra.com/

In ten weeks. Working only week-ends. Get it?

If you'll notice, that web site let's kids enter any problem and then
solve it any way they like and is able to check any step they enter
and offer hints on how to proceed, even if the software would have
done it differently. That is not the easiest functionality to
implement, so how about you spend all your time fussing with browser
variations in HTML and CSS and HTML while I do something a tad more
interesting?


I am having problems with qooxdoo? Maybe you think the focus thing was
a qooxdoo "problem"? Sorry, Charlie, my app does things with the
focus, too. Duhhhhhh. The bug was in my handling.

Cells fans will be interested to know that I got into a dataflow ccyle
where the server told qooxdoo about focus changes when the app wanted
to initialize the focus to, say, the top-left widget in a new form but
(of course) qooxdoo has to tell the app when the user clicks on a
widget to change the focus. Normally I get all hairy with code trying
to detect cycles, but this time it occurred to me to do something new
in my experience with external GUIs: the Lisp application, when it
wants to focus on X, does so by having the client eval /qooxdoo/ code
to focus on X. The lisp app then learns about the focus change in the
same way as if the user had clicked on X! Roundabout? Guilty, but
possibly Deeply Correct. We'll see if it holds up.


No, I read what they write and I can tell where they are coming
from. Especially when they end by suggesting my editor does not work
because I am using a library. See straw. Grasp.


Only to the intellectually honest observer. That I cannot help you with.


It was all a misunderstanding. I hid too well the warning that the
site was offered "as is". That is now the "front page".


Right, Kenny is the only one using JS libraries. And you call me
"fantasizing"?


Again, thx so much for bringing me back with this post days after the
last in the thread. It would have been really lame for me to try to
resurrect it myself. If you are in the city I'll buy you a drink.

The enhancements in this relase are:

- I re-ran the Algebra engine regression test and found quite a few
things to fix (and fixed them).

- In the Practice Room you can now click "Show Me" and the program
will solve a generated problem step by step. Teachers do not have to
write on the board any more, or turn their backs on the little
devils. Click on any step and the program will explain what it did.

- Miscellaneous other bug fixes.

Non-enhancements:

- No AWS instance outside the East Coast yet. Turns out one can easily
clone an instance but only within a region. I guess my software is not
the only work in progress. The next version might be presentable
enough to take to educators, then I'll worry about covering other
regions.

Known flaws:

- Many UI annoyances and confusions

- Hints and explanations refer to actual terms from the problem at
hand, even when entered by the student. But! That is only partly
implemented (it is a tad tricky, that code, and I punted a bit).

Known bizarritude:

- At some point the software will start positioning math too high
within a widget. You'll know if it happens. I have not sorted out what
triggers this, but once it starts it stays that way so once I notice
(or someone complains) I will dynamically set a flag that tells the
code to compensate. Don't you wish you were a Lisp programmer?

Next up:

- the so-called "Mission" module that offers no hints and no second
chances and requires 80% to get certified in a skill.

- "Review" lessons that offer a mix of problems from all the
subsections of a chapter. That code is there, just needs resurrection.

Cue the hounds, who have left as an exercise explaining why it bothers
them so much that a fellow programmer is having fun quickly bring a
ground-breaking application to the Web.

Again, for your convenience: http://teamalgebra.com/

kt

ps. In case anyone does not find this interesting, go to the site and
poke around and meditate on this: I write Lisp all day, and lately
nothing but Lisp. I rarely look at qooxdoo doc (tho that will fire up
again if/when I need a new widget). I almost never look at HTML and
CSS references. Yet I am building a Web application. Ommmm.... kzo

Can you please explain how to input x^2 +1 = 0

Krzysztof

 

[Pascal J. Bourguignon]

Again, for your convenience: http://teamalgebra.com/

kt

ps. In case anyone does not find this interesting, go to the site and
poke around and meditate on this: I write Lisp all day, and lately
nothing but Lisp. I rarely look at qooxdoo doc (tho that will fire up
again if/when I need a new widget). I almost never look at HTML and
CSS references. Yet I am building a Web application. Ommmm.... kzo

Sorry, perhaps I had words too strong.

Alt doesn't work. I'm using X, with a personal Xmodmap and
Firefox. (Alt and Meta aren't the same key on my keyboard, and neither
work with your web page).

What about typing sqrt instead of alt-r?

 

[Kenneth Tilton]

Sorry, perhaps I had words too strong.

Alt doesn't work. I'm using X, with a personal Xmodmap and
Firefox. (Alt and Meta aren't the same key on my keyboard, and neither
work with your web page).

Hmmm, right now ALT works for me /only/ in FireFox (my usual test
browser, meaning I neglected to test elsewhere). Investigating...

What about typing sqrt instead of alt-r?

If I spin off the math editor as a product it would make sense to
augment the editor by having a way for TeX-savvy folks to bang that in.

For kids I will eventually have a row of the half-dozen keys they'll
need, maybe even a full keypad (already executed in the desktop
version). Each key will show its keyboard equiv as a tool-tip so they
can eventually learn them.

kt

 

[Kenneth Tilton]

Hmmm, right now ALT works for me /only/ in FireFox (my usual test
browser, meaning I neglected to test elsewhere). Investigating...

Ah, I seem not to have read this until just now:

http://qooxdoo.org/documentation/0.7/keyboard_events

A little experimentation suggests a fix will be manageable.Not sure how
I missed it in testing, other than that I do not like testing. Thx for
the report.

kt

 

[Kenneth Tilton]

Can you please explain how to input x^2 +1 = 0

Type: x 2 + 1 = 0

If you want to be explicit (or type a more elaborate exponent):

x ^ 2 <right-arrow> + 1 = 0

Moving ahead, I see a problem: the editor whines about "missing root" if
one does not supply the "2" for square root (and as reported elsewhere
alt-r works to create a root only on FireFox/Windows).

kt

ps. The "Typing Tutorial" tab offers other insights.

 

[Kenneth Tilton]

Ah, I seem not to have read this until just now:

http://qooxdoo.org/documentation/0.7/keyboard_events

A little experimentation suggests a fix will be manageable.Not sure how
I missed it in testing, other than that I do not like testing. Thx for
the report.

Not so manageable, since Apple decided to put an alt key on its keyboard
but not support it. True genius. Anyway, the alt-key combos are gone and
a mini math keypad lets one get at weird characters. Not a perfect
solution, just portable. Well, too tired to test much:

http://teamalgebra.com/

Time for logins so we can do missions and record user progress.

kt