[Manifold-l] Manifold reference clients

[Dimitri Rotow]

> I think that the biggest barrier to selling hundreds of 
> millions of copies is not the price but usability.  Manifold 
> is an extremely powerful GIS tool and its use of standard 
> windows interface conventions make the program seem familiar. 
>  But as soon as you start working with it it; the 
> difficulties, without investing a lot of time, to do basic 
> things becomes apparent.
> 

I like your letter a lot, but I disagree with some of it. It's good to talk
about what a GIS that sells hundreds of millions of copies should be like.
All of us in GIS should be thinking about what it takes to expand our world
into a true, mainstream horizontal.  Won't find any argument from me about
that! :-)

Ok... to begin... I've never suggested that price is the only thing.  It is
a gateway, however, because if you price products too high it doesn't matter
what they do if hundreds of millions cannot afford them.  

Likewise, I don't suggest that any Manifold edition in its existing form is
the only product we'd propose to sell to hundreds of millions of people.  We
have the ability to alter the product with great speed, to proliferate many
different variations, to emphasize different things for different
constituencies.  It's just a question what makes sense to do in what order
for which market.

Manifold right now is aimed at a targeted user base and is designed to be
right for that user base.  It's a high end product.  We liked starting at
the high end for a variety of reasons, but one very practical reason among
many is that its much easier to do simplified variations when you start with
a fully thought-out framework than the other way around.

Another practical consideration is that we see the classic GIS companies as
being highly vulnerable to competition right now. This is one of those
golden moments when a paradigm shift can be used to push old players out of
the way.  It makes sense for us to go for a strategy that elbows ESRI,
MapInfo and similar off the desktop if that is at all possible to do, and it
is. That requires high end, serious products.

We're not in a hurry, so once we see how all that plays out we will still
have plenty of time to turn the whole planet onto GIS using whatever product
line makes sense for that.

> As an example I have listed a few of the usability problem 
> areas I have seen in the last few years (by no means exhaustive):
> 
> 1. Active columns (potentially powerful but the only 
> interface to them is scripting language -good luck attracting 
> hundreds of millions of users to scripting.  Active Columns 

This maybe a bit of a red herring.  Hundreds of millions of people don't
script no matter how you do it, but tens of millions of people who know how
to script standard Microsoft scripting languages like to be able to re-cycle
that expertise.  It's important to allow them to leverage their existing
expertise, to leverage the immense collection of educational and support
materials in standard Microsoft languages and so on.  You don't get that if
you abandon standard scripting in favor of some one-of-a-kind active-thing
language or construction tool.

> 
> 2. Basic display functions seem slow compared to Mapinfo (it 
> takes too long to pan and zoom around a data set.  I know the 
> technical reasons for this but they don't really matter; 
> hundreds of millions of users are not going to wait 10-15 
> seconds to adjust the view of their GPS track over the last 3 months)
> 

? Strange.  I work with very many fairly large data sets and I don't ever
have to wait 10-15 seconds to adjust view.  What size data are you working
with and what is the hardware and software you are using?

I disagree that the technical reasons don't matter.  If you are trying to
work with big data with old processors and 128MB of RAM, well, that matters.
Aiming at large markets means aiming for common situations and being very
good in those situations.

I do 100% agree that it's time for Manifold to expand from huge new powerful
things and also focus on the simple sex appeal of being the fastest display
and renderer no matter what is going on.  That we will do.   Although we
don't really see much competition from MapInfo (it's mostly ESRI in classic
GIS), we don't want people to say, "yeah, they are much more powerful than
ESRI but their redisplay is not as fast," we want them to say "wow! they are
*so* much faster than ESRI at redisplay, and it looks like they have more
power too."  That's underway right now.  It's a big project because doing
more than other people while being a lot faster in appearance is very
difficult, and it involves a lot more than just rendering algorithms.  But
we think we know how to do it.

> 3. Finding the distance between two or more points (probably 
> one of the more common thing people want to do with geocoded 
> datasets requires an in-depth technical knowledge of Earth 
> geometry and scripting to pull it off in Manifold. The 
> default units that are returned are usually degrees - because 
> when the kids ask how far it is from home to Disney World the 
> best answer is .98743 degrees.  I know there are good 


You can always simply use the tracker tool to get an instant read-out of
distance.  I don't know about you, but I don't use Manifold when the kids
ask how far it is to Disney World.  I use Streets and Trips, MapPoint or
something like MapQuest.  I use Manifold for GIS things and I use consumer
atlas programs when I want to do a consumer atlas thing.

Could Manifold do a consumer atlas or add an interface that builds out
consumer atlas things into Manifold?  Sure.  It's just a matter of
priorities.  For now we see greater demand for other things but I suppose
it's only a matter of time.

If we wanted to shift our focus to selling hundreds of millions of copies
would we be doing more consumery things?  Yes, absolutely. 


> technical reasons for this complexity but all the crap about 
> projection distortions is not really relevant to the hundreds 
> of millions of customers you are supposedly interested in 

I never said we were interested in hundreds of millions of customers for
Manifold System as it sits right now.  I said that price is a gateway to
selling hundreds of millions of copies and that's why I think it is good for
prices to go down in the GIS industry.  The "crap" (we should try to find a
more genteel term, next time...) as you refer to it is extremely important
to those people who do GIS who indeed take projection distortions seriously,
so it's not something that can be mishandled in a  serious GIS.

> attracting - it is good to be able to do the complex and 
> highly precise things with Manifold but you need to be able 
> to do the simple - good enough things very easily as well.)
> 

I absolutely agree with you here.  But adding a constellation of simple
things is grafting another class of product onto Manifold.  I'm not against
that, I'm just cautioning that trying to be too many different things within
one product to too many different constituencies is risky.  Perhaps it is
better to field a range of different products.  Sure, it makes sense to
expand Manifold with a reasonable range of simplified accessories but I
don't think it is wise to take that too far before it makes more sense to
continue with a different type of product.

> 4. Because Manifold can be programmed with visual basic 
> scripting it is a very powerful and flexible tool.  However, 
> unlike with microsoft office products Manifold does not offer 
> a very easy way for the average user to start using that 
> capability.  Most users start using the programming 
> capabilities of Excel by recording, viewing, editing and 
> debugging macros.  Manifold's lack of these capabilities will 
> sorely curtail its extensibility by the average user. 
> 

I srongly disagree here.  Recording macros is not something hundreds of
millions of people do, and the "average" user does not do scripting or
customization at all.  I'm not against macro recorders, I just don't think
they are the top priority.

In terms of extending the existing Manifold user base to ever more users who
are interested in programming it is much more important to extend Manifold
programming capability within classic Microsoft programming technological
frameworks: more events, .NET throughout, that sort of thing.  That's also
very important because there are many people creating custom applications
who want more powerful programming support from Manifold and doing that for
them sells many thousands of licenses at a pop. It's very high leverage.


> 5. Some of your best and most popular features are hard for 
> the casual user to find.  I think the coolest thing by far 
> with the 7.0 release was the ability to view a google map as 
> a background to my data.  This saves me literally thousands 
> of dollars in detailed street map data that I might otherwise 
> have to buy from the likes of mapinfo.  I was interested in 
> this feature and am relatively computer literate but it took 
> maybe 1.5 hours to figure out how to setup and use this  

?? It takes but minutes if one reads the Tutorial topic. 

I 100% agree with you there's a lot to Manifold and it is indeed a challenge
for the casual user to find it all.  But we don't design the product for the
casual user.  I grant you that even for non-casual users it's a challenge to
get through it all, but that's the case with any very large application that
evolves and increases as rapidly as Manifold, especially an application like
Manifold that is used in so many different ways by so many different people.
The variety of usage means there are very many possible ways to introduce
someone to Manifold depending on what their interests are.

Some people care a lot about image servers, and some people don't care at
all about them.  The range of intrests people have is breathtaking.


> feature.  A casual user would probably never get around to 
> it.  I understand that there are maybe some legal issues but 
> this is just one example of an awesome and generally useful 
> capability that it is worth making a lot easier to find and use.
> 
> Your stress on being hard is a mistake when it too narrowly 
> focuses on your current market.  You make a big deal in your 
> diatribe of letting the marketing focus be on your users and 
> having them drive your development.  But here is the problem, 
> if this list is anything to go by your customers are GIS 
> professionals and experts.  This narrow focus is going to 
> keep driving your development to provide more features that 
> appeal to the same group of users.  Your hundreds of millions 
> of copies is a pipe-dream if you keep tailoring your features 
> list to the needs of your current expert GIS market (which I 
> assume is much smaller than "hundreds of millions".
> 

Here I think you are making assumptions that are not true.  This list
represents an infinitesimal fraction of Manifold users.  The great majority
of Manifold users employ only that portion of the product that interests
them, often in simple ways (for example, they don't script, use IMS, etc.)
and they don't participate in these list or other forums.  That's fairly
typical for most applications.  Consider how many people use PhotoShop, for
example, who never participate in lists.  I've used PhotoShop for many years
and have never even visited an Internet forum of any kind for PhotoShop, let
alone posted. My experience is typical.  Yet PhotoShop is an extremely
sophisticated application, absolutely as sophisticated in its own market as
Manifold is in GIS.

In GIS as in most technical things it tends to be the wonks, like us wonks
here :-), who are on the lists.  It's not the average user.

The current focus on Manifold is mainly aimed at building the foundation we
think we need for truly revolutionary product (which we can take either up
or down market as desired) while at the same time shoving the only real
competition we see off the desktop.  And, it responds to the calls we see
from our user base.  Nothing wrong with that! :-)

> Even the feature request process that you tout is not well 
> geared to making Manifold attractive outside of the 
> professional GIS community.  The requested features with the 
> most votes are probably from the same self-selected group of 
> current Manifold users.  If you keep talking to yourself how 
> are you going to make a product relevant for people who are 
> not currently even casual GIS users? 
> 

Nothing wrong with listening to your user community.  Don't make the mistake
of assuming your conclusion.  You seem to assume it is all legacy GIS guys,
when it is not.  Manifold has an immensely broad variety of users and many
of the best ideas come from people who are not "professional GIS" folks.
That's one reason we take suggestions as they come, exactly because lists
like this are highly skewed samples that are not representative of the broad
base of users.

But, just because the average user tends to use only a portion of Manifold
in more or less repetitive ways, that doesn't mean they are simple or
unsophisticated people.  They are often highly expert within their
particular specialities, whatever they are, and they often have great
technical sophistication in things like Microsoft technologies.  In fact, it
is routine that Microsoft people coming into GIS have greater overall
technical sophistication (in things like scripting or SQL) than long-term
GIS people have.  This is OK, because each constituency brings its own
strengths, its own suggestions that end up making things better for the
overall community.  


> Your essays are always entertaining as well as informative 
> but your description of the backwards cost-consciousnnes of 
> bureaucracy while both funny and probably accurate misses the 
> point entirely.  You maybe able to sell tens of millions more 
> copies at the current price of 245 if Manifold was easier to 
> use for a wider market. Price is just one part of the picture 
> - one that you have done a glorious job with.  Power and 
> capability is another thing that Manifold does extremely 
> well.  But the final (and probably easiest) piece of the 
> puzzle is ease-of-use and that is somewhat lacking.
> 

I agree with you somewhat in this, in that ease of use can always be
improved.  There's always something, and the collective impact of thousands
of somethings can evolve things forward into truly new ways of being easier
to use.  

But on the other hand, don't underestimate the demand for raw power.
Whether you are talking fighter jets, cars, motorcycles, automatic weapons
or almost any mechanism there is always a demand for sheer power and
performance even if the price is a bit more study required for effective
use.

Likewise, the sophistication of "average" people is astonishing.  It's truly
amazing just how many people who characterize their usage as "casual" or
"inexpert" seem to leap right into dazzling use of spatial SQL and other
more sophisticated things.  I guess it's like Borland Turbo Pascal, where
the conventional wisdom was that the average PC user was too stupid to
support a population of millions of Pascal programmers.  Turned out there
are indeed tens of millions of such people, and likewise there are indeed
tens of millions of people who can do cool things with a GIS even in the
existing interface.

But for all that I still agree it's worthwhile to seek ease of use and, of
course Manifold is always encouraging suggestions in that area.

> Manifold is a great product and to make it awesome only 
> requires making its capabilities useful to a wider audience; 
> cost (at the current pretty low price of 245), I believe, may 
> be less of a factor than you seem to think. 
> 
>  

I 100% agree price is not the only factor.  But it is one of the factors
that are determinative, as affordability is a "must have" for large volume
sales.   I'm saying you need an affordable price just to be in the game,
probably in the $145 range.  I agree for truly mass volumes you also need
some tailoring and shaping of the product (or family of products).  

Plus, from a purely stylistic perspective I think it is important,
especially in the GIS industry that has so long suffered from price excess,
to push for lower prices. It's too easy to convince yourself to spend more,
to price things higher, so along the way to new and better things it is good
to always demand lower prices. Spend less, get more. Shout that from the
rooftops.  Invade GITA and tell them to start talking about prices,
especially lower prices. Go to GIS conferences and yell at everyone they
shouldn't be paying more than $245 for GIS.  Heck, yell at them $245 is too
much for GIS!  Who are we as a community to talk big numbers if we can't
motivate our own GIS peers to demand a better deal? 

Thanks for a great letter!  There is nothing better than using our
collective intelligence to figure out what it takes to sell hundreds of
millions of GIS units. :-)  More, please!

Cheers,

Dimitri