thingamy

sexy or boring - investing in the software industry

If you're an investor, there are two basic industry traits that you should look for:

1. Big market
2. Boring

Boring?

I've been an investor in many business, of which some have been of the oh-so-cool kind: Yachting, sports equipment, fashion, electronic games...

And then I've been involved in the bugger-how-boring kind: Pulp and paper, cans (yep, the fish and paint containers)...

From my many years in business I learned one thing about this dichotomy:

Cool and sexy industries attracts a steady flow of new entries, players and investors. A kind of margin influx of money that is willing to be lost and people that are in it for the wrong reason. That makes it extremely hard to earn money, if anything at all.

Boring industries are the completely opposite, the people involved stays and know what they do, the influx of new entries and money is at most a trickle. Margins are stable.

So if you want to earn money, find a boring industry!

And now I'm in software. There you have two kinds:

Consumer software: The cool and sexy stuff, especially anything web based and 2.0'ish.

Enterprise software: The boring stuff nobody sees nor gets you anywhere when striking up a conversation at a cocktail party.

Consumer software is big, but not that big. Online advertising, the income that has to be shared among the most sexy ones like Google, new media and Web 2.0 kind of stuff is approximately 20 Bn $ per year. For an indication about the rest of the consumer software market ask yourself, how much did you spend last five years on non-professional, pure consumer software (not games mind you!)?

Banks will spend about 390 Bn $ on IT next year. Banks only. Enterprisey that.
[Thanks to Dennis for the figures]

In sum:

Consumer software
Is a smallish market, has a steady influx of new entries and willing money to ensure ruined margins (work a VC packed drinks party and you'll get the drift). But that cute long legged lady at the last cocktail party was really impressed.

Enterprise software
Huge compared to consumer market, has a very restricted flow of new entries to rock the boat nor the margins much. But you need a make pretend answer to "what are you doing" when partying.

Would suggest "ski bum" to impress socially and enterprise software to impress financially. With one exception; if you're to see a VC next week, refocus your enterprise stuff to the consumer section and do not forget the colour scheme and Ajax while you're at it!

One would think that investing in enterprise software is a no-brainer, but not so, seems one actually needs brains to make no-brainer decisions. Lets hope it stays that way. Stop the calls for making enterprise software sexy, leave us alone, and bugger my lack-of-investor-interest motivated string budget.

Blink, rapid cognition, Categories and Relationships

Being out of fresh reading material over the holidays I picked up some previous read books from the bookshelves - one of which was "Blink" by Malcolm Gladwell.

Must say second-time-over-reading has it's positives - you know when you're in a boring or not so interesting part so you can happily fast-forward.

Following my recent post on Categories vs. Relationships I guess my mind was skewed when I hit a good story on some strawberry jam testing experiments:

In short, from the book - a group of food experts ranked 44 types using their usual arsenal of specific measurements like texture and taste. Taste of course split into many subcategories. Texture could be "adhesive to lips, firmness, denseness, slipperiness" and so forth. Each of those again getting a number from 1 to 15. Pretty complex, but we're talking about people with long experience and a second nature of how slippery a jam is.

Then they simplified and took the first-, eleventh-, twenty-fourth-, thirty-second-, and forty-forth-ranking jams and gave them to a group of college students. And lo and behold, their relative ranking was pretty close to the experts.

Then they repeated the same with a different group of students, but this time they gave the students a questionnaire and asked them to enumerate the reasons for the ranking. Now the results were dramatically different from the two earlier ones.

This Mr. Gladwell argues is another proof that rapid cognition or "blink" thinking is pretty good, and will be immediately ruined by too much thinking unless you're a highly trained expert.

Sounds plausible indeed, and he has lots of other examples supporting the theory. I agree, but I was lacking a good explanation as to why we're so good at rapid cognition and bad at the "long thinking".

Thus, allow me to try to add something to his theory: The third experiment was for me a classic example of using a complex system of categories or taxonomy by the untrained, a given that it does not work. Categories requires education and training. Slipperiness of jams as felt in the mouth surely more than other taxonomies - as can be witnessed by the training period for food taste experts and perfume designers, not much of four week training courses there, more like twenty years internships and much sniffing and spitting. No training and chaos ensues. As always.

But what about the second one? Or rather what's connecting the first and the second so the results becomes similar?

Taste and smell, both having very strong emotional memory (ever get flashes of memories when you pass some flowers or blossoming tree?), would give immediate relationships - "that reminds me of my grandmother's jam and sunny days during summer vacation", a sure measurement of "good taste" I would suspect.

Or the "yech, that's like the cheap stuff I have to buy on a student budget" - probably not getting high marks in the taste department.

And obviously, the whole idea about the "taste-taxonomy" would be to recognise what triggers the "good" relationships, those that triggers the good memories and the natural linking to strawberries in a field on a sunny day. Relationships thus being the common base for one and two.

If you have the book, check the other experiments and examples of rapid cognition or blink moments and you might suspect the same as me - if conscious or unconscious relationships exists in our experience we'll have a quick and easy path to decisions. Try to explain the decisions by using categories (as we're usually meant to if "thinking") and you'll find that nobody but the experts will have a chance. Obviously.

After all, relationships - between, and the actions involving, objects - is the basis for all our learning until we're put in school - and a damned good method it is. And here's good indication that categorising is fools way if you're not specifically trained so why not leave categories methods alone, accept the relationships method as better and develop it further for education and for practical daily use?

As said earlier - awast ye scurvy categories!

educators vs. enterprise and JISC

Having been to many enterprise, blogger and ideas conferences, now participating in a conference consisting of nothing (almost) but educators was a refreshing experience indeed.

As a sum-up let me list some superficial observations by a non-member of said profession:

Knowledge, learning and make meaning

"Make meaning" was a term I heard a lot. And it makes sense. Duh. Captures so much more than the usual "semantic", "knowledge" or even "search" that we usually engage in.

It could be seen as the core of what we all seek (finding the right stuff at the right time), here from an education - i.e. learning - viewpoint where it would be a core issue.

The enterprise world is slightly more skewed towards the squirrel approach to knowledge; gather and store. (Just chew on the term "knowledge management" a bit...)

A most refreshing reminder this "make meaning": Just "having" is not enough, "being" as in "assimilating it" is even more important. This I will keep close to my heart for my next enterprise knowledge management discussions! Make meaning of what we find... or make clear what we're looking for.

Best practices and business rules

Hmm, did not hear much of that. Nuff said...

Bottom up "admin" issues

Education being much less top-to-bottom run as organisations I saw a clear bottom-up need to ease everyday tasks and issues, and a willingness to try new solutions seldom seen in the enterprise world of central buying decisions and strict policies.

The need is felt everywhere, nobody likes the admin chores (who does in the enterprise world by the way?) - more time for research, teaching, mentoring and further development of pedagogic methods are where these folks wants to spend their time. Admin is hygienic and not a purpose.

I can see the wisdom in having a JISC that supports conferences like this; let the users find out and suggest, then support with funds and more. Directly.

Technology and pragmatism

Although much of the conference was highly technical, the session on Semantic Web was more technical and enlightening on the core technical and philosophical issues than anything else I ever heard. And the crowd enjoyed it and participated with gusto, perhaps slightly skewed to the philosophical side - an approach I fully support. If the underlying thinking is wrong, perfect coding and exact modelling will be moot anyway.

Still all were highly aware, and said so in comments, that although theoretical discussions are enjoyable and important, finding practical solutions that are easy to implement and allows for practical testing had to be the immediate purpose.

Typically, Scott leading the Business Process session, started out with four quadrants along the Hard-to-change/Easy-to-change x High impact/Low impact axes - pointing to the lower left quadrant - High impact + Easy-to-change as the sessions target area:
Find some processes, their drivers and impacts, that belongs in that quadrant and take it from there. And lo and behold, at the and of much fun bantering and some extremely messy sheets of notes produces by the participants he ended up with no less than 13 processes to start with. Not a bad result! Have to admit, that suits me enormously having a system with quick prototyping as a built in core feature ;)

JISC CETIS

Nope, not a conference about a new pharmaceutical molecule.

Had the pleasure of being invited to two days of discussions about IT among English educators - a first time for me, hopefully not the last! This is fun, wall to wall folks who thoroughly enjoy a good discussion, remarkably free from any entrenched viewpoints about IT.

Joint Information Systems Conference Committee - Centre for Educational Technology & Interoperability Standards. Quite a mouthful. JISC by the way is the body who allocates funds for IT all through the English higher education system.

Yesterday I had the pleasure of joining the Business Process thread (what else you might add) led by Scott Wilson of JISC.
Four hours of discussions (and I had the opportunity to bounce a beach ball around - thanks Jyri for the idea - and talk about singular objects and object-drivenness) - all expertly shepherded by Scott into a nifty action list of processes worth spending some time on, and perhaps one day build systems for.

Next few days I might be found prototyping a few of these... ;)

Logging off, have to join to the next one, Semantic Web and learning no less, goodness.

open source?

An article yesterday at CIO.com terms thingamy as Open Source. Oops...

In fact thingamy can be seen as a platform, or even a kernel - and that part is not opening it's source. For now - and here's why:

My approach is a bit like Linus Torvalds' approach to the Linux kernel: Even if it is Open Source he himself (and his kernel group) have the latest and final say as to what goes into an official release so as to assure a smoothly working GNU/Linux Open Source environment.

Those released "vanilla" kernels can then be re-built and tweaked to include the drivers you (or the distribution vendor) need.

I subscribe to that approach - a stable and standard kernel is beyond crucial as it's the basis for everything else built thereupon. It's almost like an extension of a programming language - it needs to be common for all.

Thingamy is quite like that kernel - it is a framework - a programming environment for non-programmers - to build whatever you want with it, and as such would not gain from free-for-all changes or adding of features.

A standards-based open API will allow for almost everything to be done using the kernel - a parallel to rebuilding / tweaking the Linux kernel. But the actual working business model you build using the thingamy "framework" is yours, and what you do with that is up to you - if you want to GPL a model, fine, then others can build on that under GPL. If you want to license out a neat model for the brewery industry, please do.

That means there will be a free-use "developer / testing / initial implementation" program - but we'd like to get a snippet of money when the kernel is used to run an organisation...

So for now I'll keep control over the kernel as it develops, but I do keep the parallel to the Linux kernel in mind at all times, do not want to limit any creativity out there.

Hope you're all OK with that for now :)

The Semantic Web will not happen. Yet.

There, I said it. Stuck my head out precariously. Feel free to chop it off, but in a nice way please.

Although, the "Semantic" part of the term can indeed be used, and that thingamy intends to do. But it won't work for the "Web" (as in the whole Cloud, the WWW), for now.

The cloud (web) holds information and the information is contained in (web) objects - pretty much unstructured and rather messy. Not a problem, just something we'd like to sort out, Semantic Web wise for example:

Applying Semantic Web methodology would include adding RDF N-triples - [subject] + predicator + [object] - to web objects. Computers understand it (thanks to W3C standards) and humans get it, perfect.

But, there's the quandary; each web object (mostly) represents a multitude of real world objects. Each information bearing object - article, web site, form - is messy with a multitude of object representations within itself.

A form represents a person, an account, an address and many other real world objects. An article is a narrative describing many objects and referring to many more. Both have owners and authors and purposes - yet more, virtual or tangible, singular real world objects represented therein.

So which part of the subject and object would the predicator apply to? What is actually related to what?

In order to add precise relationships in the cloud, web objects need to be singular, each representing one single real or virtual object. Precise objects.

It is all about knowledge. How objects relates to other objects equals knowledge. Not more, nor less, that's it. Knowledge in a nutshell.

And without precise objects, no relationships can be precise and thus little knowledge created. And therefore no successful Semantic Web (all of the WWW that is) for awhile.

(BTW, narratives and forms used to be the solution to hold knowledge as it includes all the important relationships therein. Neat method in the paper days, but now when we'd be better off having something that computers can understand? Nah.)

What to do? The idea of a Semantic Web is too good to let lie.

By now you may have guessed it, use it only for pockets of web based information that fits the bill of object singularity. That's what I will do in thingamy:

And we're extremely lucky, we operate in a controlled environment, in an organisation or business, where you not only use information but you also create the information - thus you can define the classes such that the objects become truly singular.

Then the triple offers precise relationships. And with precise relationships between precise objects, knowledge ensues.
That computer readable knowledge allows a system to find, use and report objects according to any relationship. Real time processes, Semantic Accounting (!), reports combining objects or drilled down, up, sideways or wherever you want to go.

Semantic Enterprise before Semantic Web?

No more tags!

Seriously. Wrong path. Blind alley. Tags are out. Full stop.

Yeah, yeah, I know I've been a strong proponent of tags - but now I accept I was all wrong, very wrong.

Tag me with "France" - would that entail my passport is French? That I live in France? That I like the French Alps? (Two last applies)

Tagging subjects is like ripping the verb out of a sentence.

"Sig France", eh?

Put the verb back!

Try for example "Sig + lives in + France" which technically is "subject + predicator + object" which, for those of you who are conversant in the "Semantic Web" is similar to W3C's RDF N-Triples (heh, how's that for a mouthful).

In other words, a methodology and even a technology exists already that "puts the verb back". So why not use that?

This framework has been around since 1999, with a new version since 2004 - but still it has not caught on despite it's intriguing prospects.

I would dare to suggest that this non-use is due to how information is stored:

The free-forming cloud, the "web", as well as enterprises still handle information in mashed up formats - "documents" and "forms". Storytelling. Mashups.

A "document" almost always consist of many objects - address, author, many subjects and even more objects - like a vat of 12 liquids nicely mixed up.

Add a triple saying "vat contains inflammable liquid" - which one of the 12 liquids does that apply to? If the information had been stored as 13 singular objects, 12 liquids + the mix, then triples could add real useful knowledge.

Luckily thingamy uses only singular and minimal objects to build models of reality. Thus triples can be used with aplomb.

Out goes tags, owner, place and links - our current "object knowledge enhancers" - to be replaced by one single method; Relations (triples) - and with that; one single term to get head around, no limits to knowledge enhancement, real precision and less code.

So apologies (again) to those of you keen to put your hands on something (almost) finished, just have to get it right when I see a flaw.
Bugger, it's not easy to be "on time" when you're breaking all the "rules", then breaking your own rules.

More on how we're doing it and how it's going to make thingamy simpler and more powerful than ever in post to come.

death of the Applications

That was the title for a panel at Office 2.0 in SF yesterday, you'll find it here.

Well, here's my wee little 1-2-3 twist to that issue:

1. Once we invented tools to help us in our tasks; hammers, knives, drills, machinery, cars, typewriters...

2. Then IT created new versions of many tools - the "applications": Word processors, accounting systems, e-mail, spreadsheets, new communication tools, social software...

3. Then somebody says "Hey, IT is about modelling reality, so why don't we model the whole flows instead of only the tasks?". And with that the task-completing-functionality becomes a natural part of something holistic making most stand-alone task tools irrelevant.

So be it, holistic modelling making "applications" redundant.

(And with that the end of middleware, never ending discussions about file standards, SOA and other architectural approaches to the complexity created by the application-jumble.)

Storytelling and running enterprises

I like storytelling, I like Spielberg and JK Rowling. The conflicts, the suspense, the events and transactions enhanced with what people think and feel. All tweaked and bent so the story flows better.

Just like accounting - summarised reports of events and transactions, all tweaked so the intended story is easier to follow.

That's the beauty of using events to paint a picture, lots of freedom but alas little reality.

If you're working in an enterprise, when you order a widget you describe the first event with an "order sheet", then the shipping event with "shipping papers" and so forth. To understand the widget, or understand Harry Potter, you have to study the events and put them all together, a whole stack of reports and sheets in the enterprise world, or a book in Rowling's world. But just as for Harry Potter one event might throw you off and conflict ensues. In accounting that's called reconciliation.
And of course, add SOX and other regulation rules to keep a lid on things.

This is how enterprise software is designed, event documentation, or as some would call it, storytelling. Even the enterprise tools like word-processors and spreadsheets are support tools for the storytelling, and worse, made for story manipulation.

This because event documentation, or story telling, was the only way to capture the essence of the enterprise when we only had pen and papers. Event-driven and event-documentation. Or transaction-driven - same concept but for stuff with value.

But there is another way. It's called object-driven and object-representation.

If I'm at a football game my eyes follows the ball and registers it's every move, the spin, the direction, the goals. The game is object-driven. If the ball looses speed and stops up I can deduce nobody has kicked it, the object represents what happens. Obviously I'll also keep an eye on the other "objects"; the ball-kicking-assignees - the players.

The plants in my garden "tells" me if they get too little water or sun by signs of dry leaves. The plant-object captures all that happens to it and I can deduce what happened. I'm using "rules" and "mapping templates" in my studies of the object - dry leaves, short of water - then I'll increase watering by changing the "aperture" property of the "sprinkler" object.

When I'm driving, my eyes are peeled on other objects; cars, people and cyclist - objects with telling signs of what's going to happen is "told" by the objects themselves. No doubt also a object-driven environment.

The assembly line used to be like that, the car arrived without doors - I have doors at the ready and tools to fasten them. No need for a "work order" I can deduce from the objects themselves what my task is, and when. After that the object itself have captured what has happened to it, no need for an event report as the doors are visibly in place. If they were not the object itself would have "told" that I fell short of my task.

This is possible to attain using IT: Let any real-world object be represented by one data-object, punt those data-objects letting the real objects follow along the sequence of work that shall happen in lieu of work-orders. Then let the data-objects capture what happens and use templates with rules to produce any report with immediate effect.

What about virtual objects? No problem, replace the "car" object with a data object representing say an "issue" or a "medical condition" and add details as properties of such data-objects and punt it to an expert or the doctor who then decides where to punt it onwards. X-ray perhaps? Another expert? The work-order or appointment comes as a fully displayed data-object just like the car on the assembly line, no doubt what to do, no need to tell "please fix this patient" and data capture happens through the object properties being filled out or changed as well through capture of who and when did it all.

That way an object-driven model of the enterprise will open for far easier modelling of barely repeatable processes as well as the classic linear ones, have much built in run-flexibility and deliver reports real-time in any format with no need to worry about reconciliation. SOX and other rules all covered by design too.

Boring perhaps, storytelling being so much fun - but your shareholders will thank you and the CPA will have less to do.

merely social software or objectively social

I always enjoy listening to Jyri Engeström, and this year again he did not let us down.

He was looking into social software services of the kind we all link to each other in these days - LinkedIn, Facebook, Flickr, YouTube, MySpace...

Sitting at the back of the auditorium my mind was on all the places where I am and where people link to me and vice versa, often the same people in more than one place. In the beginning kind of fun, but what then?

Jyri's postulate was that services offering nothing but social connection - LinkedIn was mentioned - will have a limited life span. Some kind of social object has to be a part of the experience according to Jyri - like photos, videos, or at minimum some of the music objects that Myspace focus on.

And I tend to agree, all the linking is nice enough but the "so what" feeling is there already.