I wonder if you'll change your view on this as you age.

I'm 40. I've been through 2 divorces and 3 full-on professional careers (academics, music, and programming). I have a 17-year-old son.

Although I can probably think out the evolution of my "software", it's really hard to think of my current releases as having a whole lot of backwards compatibility with that 0.0.1 version. If I had to give it a number, I'd say I'm somewhere around release 6.4.3.

:D

Personally (and this is just my conclusions, not one I expect other folks to come to), my views and strategies for working with the world are so removed from 25-year-old me that I have little problem saying that I am almost not the same person.

Certainly, my "software" has so many revisions to essential functionality that it's hardly recognizable as the same collection of code: there is very little overlap between how I reasoned about things when I was 25 and now, and I feel like that's not just a matter of "data" aggregation. My conclusions is that the software isn't, in fact, the same.

I feel like it's more that I've fundamentally revised the software on a number of levels to the point where there are so many essential differences the identity isn't even consistent.

I agree but for me the business logic is the same. It's trying to solve the same problems. I want the same things I always wanted and I'm not talking about money or material or social status and relationships. I still want the same specific skills and experiences that did I when I was a kid but outlook is totally different.

> The data is various, but the software is the same. The consistency of reasoning between now and then is incredible.

Don't underestimate the environmental factor that your journal provides. Perhaps you're most consistent than most because you can and have read your prior journal entries, and this preserves the consistency of your reasoning over time.

Kind of like studying: it's well known that more frequent review of material dramatically improves retention.

I agree, I've actually made a point to not read older entries before I write anything new in mine and wait until afterwards or an entirely different time specifically for reading it and not writing because I've noticed this effect in myself.

Things you write down into your journal have already been under psychological screening mechanism described in the article. You don't write down things that don't seem relevant to you or things you don't remember when writing it down. You are creating personal narrative and documenting it.

Another screening happens when you are reading your own journal. We are not objective readers and we add and remove things into what we read.

Your case is somewhat particular (and is a very good thing to do). Because you have a log of past events, your memory processes can match against a reference and filter out deviations.

Most people don't have such an external log/reference. Memories cannot be externally validated, only internally validated against other memories. So, small deviations can accumulate, and overtime create entirely false narratives (esp around uncomfortable or even traumatic events)

You make me want to start a diary :-)

Our software wouldn't be very interesting if it wouldn't be able to at least partially modify itself.

But that's a very interesting experiment you are running. I wonder if you will still hold the same conclusion after next 20 years. Although, you can always say that the data is being interpreted by the software and effectively say that the source code changes but the interpreter stays the same.

I think it's more like a Moore state-machine, coupled with some invariant functions. The events come in, and they alter the internal state somewhat thereby altering future outputs elicited by similar events, but the invariant functions mean that certain things will never change over time. Of course, as with any real-world system at operational limits unpredictable behaviour can emerge.