[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: st: significant digits format
David Souther <firstname.lastname@example.org>
Re: st: significant digits format
Tue, 12 Jan 2010 11:52:38 -0600
On Tue, Jan 12, 2010 at 10:19 AM, Roy Wada <email@example.com> wrote:
> No, Ben. You are acting like a student who gets caught and complains
> that he did not understand that paraphrasing someone's work without
> citation is plagiarism.
> Yours: local left = int(log10(abs(`value'))+1) // digits before dp
> Mine: local left=int(ln(abs(`input'))/ln(10)+1)
At some level every program is probably inspired by some part of other
Both of these look exactly like the variations of formulas I've seen
used in Excel or VB script to get at this same issue (though the excel
version looks more like
"=ROUND(value,sigfigs-(1+INT(LOG10(ABS(value))))" ). A google search
of a formula for significant digits turns up both of these approaches
in Excel & VB (as I imagine it would in other stats packages). I
think if you compare packages with similar functionality in Stata,
there will be lots of similar approaches to create that same
functionality--there are really only so many ways that you can do
things in Stata (at least, do them efficiently).
Ben has acknowledged that his programs were influenced by
outreg/outreg2 (in fact, it says it was 'stolen' (flippantly, I
imagine) in the .ado file--though I have no idea if this
acknowledgement came after some prodding by Roy), and Ben has
indicated in other postings that he was building upon existing
knowledge. That is something that is desirable and accepted in
academic work; what's the benefit of starting from scratch every time?
If Ben wanted to create an outreg-like program with a different
syntax approach & different goals for its functionality that he found
useful, why should he be expected to reinvent the wheel blindly in
order to arrive at something with overlapping functionality ? Though
I'm not an expert on either package, I tend to prefer the syntax and
output that esttab/estout creates, I'm glad that these packages were
What's more, I haven't seen any guidelines on Stata or SSC (or in
open-source programming in general) that defines how authors must cite
others' open-source work or that prohibits someone from using other
code to help recreate functionality. I don't see that you have any
kind of creative commons or copyright on your adofile.
> By your own words, this is a "stolen" work. This makes it a
> plagiarized work. And that makes you a plagiarist. This is
> unprofessional and academically dishonest. It needs to stop.
> The prattle about being similar or not similar is beneath you.
> -sigdig- is a rewriting of SignificantDigits. I don't keep tabs on
> what you do. I take people at their word, but your words fail me.
> You are fooling yourself and many others if you say that you are
> making useful contributions by deliberately knocking off other
> people's work to the point of extracting codes and inserting them into
> yours, putting your name on top, and misrepresenting yourself as the
> contributing author.
> It seems you are unable to correct these errors on your own. Based on
> the earlier post, you seem to think there was nothing wrong. Listen to
> me carefully because you don't seem to appreciate the enormity of what
> you have done. You have a history of taking other people's work and
> misrepresenting yourself as the orignal author.
> -esttab- is a plagiarized work. It is based largely on -outreg2-. This
> was something that you could not have written on your own. It would be
> nice to if you could have, but based on your other work this was
> apparently something that was beyond you. What you did was blatant and
> shameful. So shameful in fact that I don't know why you are making me
> come out and say these things.
> -estout- is loosely base on the original -outreg-. By distributing the
> program without giving adequate credit to the original source (not
> given in the help file, not in the website, not the handouts and
> presentations floating around), you have plagiarized John's work. This
> has greatly hurt John's reputation, giving rise to the impression that
> there was something wrong with his work.
Thus far, you've presented two similar lines for a common approach to
a formula as your evidence. I think these kinds of accusations demand
that you give us more. Where's the line-by-line or
subroutine-by-subroutine comparisons of code that definitively prove
these statements? If you really think the code is that similar, file a
> This was completely untrue and highly unfair. Unlike you, John was
> highly original. I admire originality. I don't expect you to share my
> values, but I do have problems with people who do not understand any
> of this yet wishes to present themselves as a scholar. Scholar you are
> not. Scholars respect the integrity of other's work. Author you are
> not. Authorship does not consist of "stolen" work.
I think the Stata community is pretty lucky that authors like Roy,
Ben, and John would create these useful and important packages, but I
also think we are lucky that Stata allows a forum for us to exchange
and share these add-ons free of charge. Not all stats software
companies allow this sort of thing. In this sort of environment, I
expect that there will be similar/redundant packages that come about
and it is up to the user to decide which package works best for them.
* For searches and help try: