I get a little bit disturbed about this Wolfram Market page about Debugger , and create this reply thread on Wolfram Community. Comments are welcome. I believe that Debugger is our next CTRL+Z. :)

@Murta great post. Much better than my temper tantrums when I see this kind of thing. Very constructive.

@SigismondKmiecik Hi, sorry to miss you. I'm rather under time-pressure at work to finish a project. I've been working all weekend, aside from checking in to procrastinate. :) I was thinking it might be appropriate to ask your Q about fline on the main site, where you could post the code, too. Someone might be able to help. I guess you're making some sort of mistake and there are many people who can help.

Hi folks, can you give me a feedback whether I came across too harsh here:

mathematica.stackexchange.com/questions/47683/…

Seems like my comments did not go down well with the OP.

I got a somewhat conspicuous downvote shortly afterwards here

@YvesKlett This sometimes happen for all of us and you apologised.

@halirutan so you think it was too harsh?

@YvesKlett That was me. I thought it was the right moment for tactical downvoting!

@halirutan hehe you old agitator :D

I guess I will bear the scars with pride.

@YvesKlett No, I don't think it was too harsh. It clearly stated how the question should be improved. I guess the OP is a bit zart besaited..

I understand that his circle-packing question got great answers, but I am not an immediate fan of just pasting wish-lists.

@Murta I'm relieved to read your comment on the Debugger as I never had success with it myself.As I'm a rather modest MMA user I thought that my inexperience was at the core of it and never did I suspect that this facility is not reliable. It also explains why I never dared to ask a question on SE about it for fear of being downvoted.

If you read [here](mathematica.stackexchange.com/questions/47221/… ) I have a problem with a module that does not work when called from a dynamic context (otherwise perfectly OK, I had used used it quite often).

I'm pretting sure that it is something obvious ,nevetheless I spent hours in vain and the Debugger is hopeless to me. Asking then again a question on SE following an already nicely answered question by @Michael E2 I'm reluctant to do. Can anyone on the chat pinpoint to me where the error is?

@YvesKlett harsh ?! not at all. Plus, he's been an user for 4 months .. That's just plain lazy/lame

@YvesKlett To somewhat relieve your pain: 4coloring.wordpress.com/2011/04/25/… math.stackexchange.com/q/39510/2731 mathoverflow.net/q/63861/12700 and this last one posted almost simultaneously (I'm not sure if it's related, though) mathematica.stackexchange.com/q/47701/193

@YvesKlett I used Google search-by-image. Those graphs were too professional for such an odd question

@belisarius wicked.

@YvesKlett That comes with age. Wait and your turn shall come

@belisarius truly sage you are!

@YvesKlett I prefer "maven" if you don't mind

(and thus you really go well with saltim bocca)

@belisarius uuuhh - nice, my daily fresh word.

Silly association: Mavenrick

@YvesKlett :)

Up there with the best of the best :D

@YvesKlett The original planar graph to circle packing algorithm paper is linked here: math.utk.edu/~kens

@MichaelHale thanks for the link!

Currently I am doing if = Interpolation[cdf, InterpolationOrder -> 1]; pd = ProbabilityDistribution[{"CDF", if[x]}, {x, 0, 3}]. Then I can use N[Variance[pd]] or N[Mean[pd]], but not N[Median[pd]].

@Szabolcs How is the CDF between the points?

Smooth, constant?

@Rojo Let's say it's linear.

@Rojo So the PDF is a step function like this:

There are so many built-in distributions that it is too easy to miss some of them. I was wondering if I did miss one which is a good fit.

@Szabolcs Perpahs HistogramDistribution with WeightedData

?

Looks interesting though it seems to take actual data points rather than PDF/CDF values. I'll take a more detailed look.

@Szabolcs Doesn't seem to fit perfectly

but it has to be some DataDistribution

The thing is that DataDistribution is not documented so, to generate it manually would mean to spelunk it a little bit

@Rojo Yes, you're right. So it looks like DataDsitribution["type", data, something, dataLength]. I'm not sure what the something parameter does. The data part looks clear for a HistogramDistribution.

@Szabolcs I'm guessing it's the dimension

Not sure

@Rojo Correct!

Now, where would the last parameter be used for a HistogramDistribution? It is the number of original data points. It is not present in the histogram data embedded in the DataDistribution.

I mean, what I want to do is create a HistogramDistribution from actual histogram data, not data points. Can I just make up the number of data points?

@Szabolcs I guess there's no problem

as long as you don't ask for the property "SampleSize"

@OleksandrR. About a month ago I went to a WRI demonstration at our university. To me, it wasn't very convincing because it showed very little of the real power of Mathematica that I actually make use of. They showed a lot more of the showy/pretty stuff (such as presentation functionality, predictive interface, natural language queries, interactive graphics editor, built-in units, 2D math notation) that I never use in practice, and I don't find very functional.

However, looking at the people who attended, most of them were quite old, and they didn't look like people who would use Mathematica. They looked like people who would make decisions about buying it ... that explains a lot of the stupid and aggressive marketing (which only has a repulsive effect on people like us who actually use the system)

@Szabolcs FOr your data, perhaps

@Rojo Yes, something like that. I want to post this as a question. Do you want to post an answer yourself or should I post it (if you don't want to take the time)?

@Gabriel Nice that you liked it. The idea is to be construtive!

@Szabolcs It's ok, you post it

@SigismondKmiecik I agree, a lot of people think that the problem is that they don't know how to use the tool, and not the tool itself. I believe that the problem is that the tool do not work.

@szabolcs also spliced distribution with weighted uniform distribution. Sorry typing on phone.

@MichaelHale Thanks!

@Szabolcs

Plot[PDF[SplicedDistribution[#[[2 ;;, 2]], #[[All, 1]],
     Table[UniformDistribution[{Min@#, Max@#} &@#[[All,
          1]]], {Length@# - 1}]], x], {x, -5, 5}] &@{{-1, 0}, {0,
   1}, {1, 3}, {2, 2}}

@Murta I don't usually use it but when I've used it, I didn't find it sooo buggy.

Do you have examples apart from the Abort one?

@Rojo I have to jump to Murta's help here. I don't use the debugger but I have tried it and setting break-points which work and debugging code is a pain in the ass. I really want to see someone using it with 1k lines of code.

@Rojo Btw, completely unrelated: When AbsoluteTime gives the number of seconds since 1900, why is it a Real?

@halirutan I don't say it works nicely, or that it is bugless. Just that I heard lots of complaints and few concrete examples, so it's hard to try and jump in

@halirutan AbsoluteTime[] is always accurate down to a granularity of $TimeUnit seconds, but on many systems is much more accurate.

@Rojo Yes, this is right. I think a lot of people just gave up on it. Maybe one should really try to use it and work out some specific problem examples.

@halirutan Exactly

@MichaelHale Thanks (It was more a rhetorical question :-)

@Rojo in this post I show a concrete example. The abort button don't work. The same happens to run to selection and watch point.

I have always heard complains about Debugger, so I tried test it myself. And yes, complains make sense.

@Murta Yes, I saw the abort example

Does that example work to show the issues with Run To Selection and etc too?

The abort thing is clearly buggy, but you can always abort "manually"

I guess