[ocaml-biz] Replies to lots of messages...

William D. Neumann wneumann
Thu Aug 26 12:17:34 PDT 2004


Well, a power outage at the university has made my mail access fairly spotty for the past few days, so 
what follows here is a digest of responses to messages of the past few days:

> I'm amused that such a rock exists. ?However, it doesn't have worldwide 
> identity for brand recognition. ?It's not like everyone in the world 
> knows about this rock, it ain't no Rock Of Gibraltar or whatever. 

And to that I say, "So?"  Quick:  What mountain is used in the Paramount logo?  And I'd bet that if you 
asked people what rock is pictured in the Prudential logo you'd get as many "Fraggle Rock" or 
"Monkey Island" answers as you would Gibraltar.  The point is you don't need people to know what 
the hell your logo is, so long as it projects the right image to the right people.  Now I'm not saying 
that Camel Rock is the right image for the logo (while the idea of a rock formation might promote 
safety and stability, it also tends to project a ponderous, slow image as well), just that familiarity 
with a given object is so much less important than the looks of the logo and the qualities it projects.

> Maybe 'robust' is better than 'safe'. ?To me, 'safe' has the connotation 
> of 'boring'. ?Maybe 'high performance' is a better mantra than 'fast'. 
> Just sounds flashier. 
> So: 
>
> ? ?high level, high performance, robust, pragmatic 

Well, you seem to be discussing a slogan or pitch for OCaml, which is pretty much an orthogonal 
problem to the design of a logo.  I've been doing a lot of looking at logos lately, and really, quite a 
few of them have zero (or close to zero) connection to the products or companies they represent.  
For example, the following logos say very little about the products they represent:

Apple, IBM, Dell, the O'Reilly Perl Camel, NBC's peacock (while it once had significance back when 
color TV was in it's infancy, it no longer has any meaning but a historical one), pretty much every car 
manufacturer out there (notable exceptions are the Dodge Ram and Ferrari logos), most 
airlines, PHP, MySQL, just about every shoe company (with the slight exception of Nike), Red Hat, 
Debian, Sun, Coke, Pepsi, etc...

Just a reminder that descriptions of OCaml can be saved for the slogan "Quality Is Job One", "Think 
different", "Zoom Zoom", etc. but it need not be expressed in the logo.

> I hate to say this, but in today's political climate of irrational fear 
> and hysteria, picking as our mascot a person of middle-eastern descent 
> would be a PR diaster.

To a certain extent I (sadly) agree.  However, this doesn't mean that all middle-eastern images are 
verboten.  One could easily get away with, say, Aladdin on a flying carpet, a Genie (if it weren't 
already taken by Genie), or even date palms with minimal fuss -- though I don't think those are 
exactly great ideas myself...

> 'Automated' might be a good marketing buzzword. ?It's not literally 
> true, and people will complain, but a certain amount of complaining is 
> actually good in the "all news is good news" marketing hype sense. ?What 
> I like about the word 'automated' is it emphasizes programmer time, as 
> compared to 'efficient' which really doesn't say anything about how long 
> it'll take to get things done. ?ASM is efficient, after all. 

Not a fan of "automated" here.  I can't even get a good picture of what it would mean in terms of a 
(real world, not fantasy) programming language.  I understand that the image of the PHB is pretty 
popular (though I've never actually met a manager like that in my life -- I guess I'm just lucky), but I 
can't imagine a great deal of suits in tech oriented businesses (our primary target, right?) believing 
that there's this unknown language out there that you just have to tell it what you want and it spits 
out all the code for you.  it almost strikes me as absurd as John Deere using "Bootylicious" as their 
pitch (not quite as absurd, but close).

> 1) is this even vaguely appropriate for selling a language technology? 
> Usually when techies say a tech is 'sexy', they don't mean it's 
> literally a woman with nice gams and so forth. ?They usually mean it 
> more in the sense of a hotrod or well oiled machine is 'sexy'. 

Is it vaguely appropriate?  It could be.  I'll be damned if I can think of a logo that I'd classify as sexy, 
but I wouldn't rule out a logo just because it's sexy.

>?I find the connotation of 'mysterious' to be completely negative. ?'Alluring' 
> is ok, but it has to be the right kind of lure.

I don't know -- we might have a different idea of what mysterious is, but I'd definitely classify the 
Red Hat logo as mysterious.  And the AT&T death star logo as well (to a lesser extent).

> I don't think they are similar enough to 
> the Perl's camel to cause any legal problem with OReilly (we still need 
> to ask official permission if this proposals go mainstream). 

I don't know about that -- to me they look like a slightly cartoonish version of the O'Reilly camel.  I'm 
pretty sure they'd feel the same way about it.

> To say more, I'd have to get out my sketchbook and try some things. ?I'm 
> not quite ready to do that yet, I'm still in "define the message" mode. 
> That said, I envision a black 'O' with a camel as the whitespace, the 
> center of the 'O'. 

And I say break out the sketchpad for three reasons:
  1)  As I covered above, logos need not say anything about the product they represent.  They need to 
look sharp and positive, that is all.  If you can get a message in there as well, that's a bonus -- but 
it's hardly necessary.
  2)  You can try to get the message "right" from now till doomsday without ever actually doing 
anything.  The OODA loop sort of applies here -- superior analysis and no action is a losing strategy.  
Analyze, act, refine in the next iteration of the loop.
  3)  Words are lousy conveyors of information about images.  We all know the old cliche about 
pictures and thousands of words -- it's true.  Plus, if you put up an image, it will allow others to riff 
off of that image and move the process along.  Nothing inspires like inspiration -- or something to 
that effect.

> First of all, three years of art classes in High School has convinced me I 
> have 0 artistic talent. ?So don't expect sample art from me, 'cause I 
> suck.

So?  Have you seen the crap I've been throwing up there?  At the very least, do a google image search 
to find candidate watches or cogs or whatever floats your boat and put them up somewhere to say, 
"See... I'm thinking something along these lines."  You say "swiss pocket watch" and I say, how is that 
different from your average pocket watch... show me what you're talking about.

> I would definitely say, design a proper business logo *first*, have a 
> plushy toy mascot contest *later*. ?The two should not be confused for 
> each other.

Well, they could be done simultaneously, I don't see any harm in that (other than perhaps diverting a 
bit of energy from the logo effort).  However, you are absolutely correct when you say that the two 
should not be confused (and really don't have to have anything to do with each other).

> The logo is far more important than the mascot.

I wouldn't say that.  Techie mindshare is pretty important as well.  It's better to make a double 
fronted effort, one that appeals to techies and suits (through different means, of course -- doing 
both with one package is a *hard* task) is better than one that appeals to just one or the other 
(though if we do go just one way, I agree that suits come first).  I guess I'm saying that we shouldn't 
just toss away good mascot candidates -- they might be as effective a byproduct of the logo making 
process as whey is from the cheesemaking process.

> Who do you think got the Linux marketing job done? ?Tux or Redhat? 
> Actually, I'm not fully conversant with the history of Linux marketing, 
> as I dropped out of the Linux world in 1996. ?If anyone remembers the 
> history of what paved the way, we'd welcome the comments. ?My impression 
> is that *much* is owed to Redhat. 

I'd say it was the combo of Red Hat and VA Linux.  However, I wouldn't really say that it was anything 
involving Linux as a *product* that those two brought to the table, but rather they were the first 
groups to show that a viable business model can be built around open source/free software.  It's not 
that Linux the product had a bad rep with suits, it's just that none of them were creative enough to 
think of a way to make money off of it before RH & VA.

> If Ocaml becomes popular in industry, this natural pressure will also make 
> it popular in academia.

s/popular/prevalent/g and I'll agree with that statement.  I've been involved with academia and 
academics for the past <mumbledy-mumble> years, and in all that time, I think I've met maybe five 
people who actually like languages like C/C++/Java.   For the most part, they teach them only 
because they *have* to (which I suppose is what you were saying).  It's really an "Are we teaching 
programming, or computer science" kind of thing.  

But you're right, the bottom line is suits > techies >> educators.

BTW: A couple of language logos I like can be seen at:
<http://www.jeda.org/jeda_logo.jpg>
<http://www.seas.gwu.edu/~mfeldman/ada-educ-logo.jpg> (not a fan of the text around the logo, 
but I think it's a good logo.
<http://www.heron-language.com/>

William D. Neumann



More information about the Ocaml-biz mailing list