Due to my experience it's quite easy to understand the difference, though for me the difference wasn't obvious for long time.
Let's consider yourself sitting on a chair in your kitchen looking at a table. On the table there are 10 apples of different sizes of two sorts: 3 apples of one sort and 7 for another. Also there are 3 knifes, 2 lemons, and 4 credit cards.
Obviously, we can identify concepts or types like: Apple, Lemon, Knife, Credit Card, Kitchen. We can go from generalized concept "Apple" down to subconcepts of "Apple" of each particular sort, giving definitions to recognize them in real world with our senses. It's obvious that every Apple, every Apple is unique, has identity if we try to compare them precisely, - we'll find they are different. Right?
And this is Conceptual model or Type model or Domain Model where real objects play major role. This model provides structure for natural language used to talk about problem domain . And your above definition is right. We IT guys and analysts require it to be build precisely to get rid of ambiguity, errors, fictions, misunderstanding, slang and etc we normally face on every project where there is no terminology or it's outdated or all the knowledge is in peoples mind and not formalized.
As we are sure with what things/concepts we deal in a problem domain and how they are related, based on the automation task specifics, we can design proper information model. Goal of information model - is to represent structure for collecting information about concepts required for task being automated. Information model can be represented as sheet of paper or database. That's not important. It's just about structure of records about reality required for specific task to be automated.
Can be represented as lines of text on a paper:
- 3 apples of sort X on table #1 in kitchen #1 at location #1
- 4 apples of sort Y on table #1 in kitchen #1 at location #1
You get rid of each apple identity and record quantities of things of a kind, sort at specific location and etc. Technologically we cannot record every apple as unique, because there is no way to get and record it's identity. Even if we find and record 100 apple characteristics, theoretically there can be two identical with all 100 properties the same. In reality we have an option to find characteristic #101 to differentiate, but information model is mostly static (AI changes this sure. I talk about std information systems). So you have to quantify things.
But with credit cards situation seems different, but generally the same:
- 1 Credit card #1232233 holder DDDD in kitchen #1 at location #1
- 1 Credit card #2334234 holder FFFF in kitchen #1 at location #1
We rely on card number pretending it's unique. However due errors theoretically and practically there can be 2 cards with the same number and holder and other properties at the same location. So in information model you'll also record it's quantity if you really want to record reality.
So, the main difference between conceptual/type model and information model is that in inf. model there is strictly saying no possibility to talk about object identities at all.
PS: And what about Primary Keys, Unique Constrains and so on? They are all artificial. In other words you can constraint your info model/database to have just one credit card of each number. But in case someone will issue the same credit card or duplicate and insert it into the same ATM your where near 10 minutes before with your original card, how will you recognize it in information system? No way.
But if two "identical" credit cards are lying at the table before us, we can try to differentiate them by doing precise and long comparison. We can thing and talk about each card as about different objects even without knowing where they are different. It is de-facto assumed in conceptual/domain/type model.
Hope this will help to understand such tricky difference.
- Definition of Information model you've given I think is correct, but not the best.
- Axiomatically someone can argue that there is a possibility to create two completely identical objects... Maybe it's possible. Personally I don't know how to employ the same atoms and electrons for two different things looking completely identical.