Hacker News Reader

Comment by jdgoesmarching | Hacker News Reader

Do we need to have a moral panic every time we ascend a level of abstraction? Hierarchical folder structures are mostly for the benefit of human organization, and over the last decade we have found different ways to present information such that the average user of newer tech doesn’t encounter it.

Recently I’ve be de-foldering a lot of my digital life and getting by with tags or more powerful search tools. I wonder if it would be more worthwhile if we started teaching basic querying methodologies rather than worry about underlying file systems. The next generation will grow up with far more data than can be easily traversed in folders.

View on HN

>The next generation will grow up with far more data than can be easily traversed in folders.

"If you can't find it, you don't own it" - Mike Warot

I think the real lesson here is that there is a cost to hoarding data instead of curating it. The reason files and folders were brought into use in the first place was to store masses of data in a way that allowed rapid access, that avoided search. Huge amounts of labor was spent, in very small increments, to make this happen.

< tangent > The filing system that indexed Marriage records in my home state of Indiana for the century prior to computers was quite elegant in it's efficiency. Marriages are recorded sequentially, in sequential books. At the start of each book are a number of alphabetically organized pages with space for names starting with A, B, C, ... Z with space allocated roughly matching the frequency of the names, but twice as many listings as Marriages. The marriage was recorded in the next available entry. The record number was then placed in the index, twice... under both combinations of last names. Normally, you could find a Marriage record in a few seconds, worst case if you didn't have the date, it's sequential searching through the book indices. If there was an error in the record number, you could check the other entry as index records had both names. Worst case scenario was a sequential search. < /tangent >

Just because the computer has superhuman speed at searching for text, if you can't describe something, you no longer effectively own it, if you have a hoarding situation.

Just like all those clerks in the centuries before us, if you want quick retrieval, you have to put in a bit of effort and planning ahead of time.

I think the real lesson here is that there is a cost to hoarding data instead of curating it. Just because the computer has superhuman speed at searching for text, if you can't describe something, you no longer effectively own it, if you have a hoarding situation.

We need to better understand the trade-offs ourselves, and teach them to our successors.

As someone who's been in this kind of position, I don't see it as a "moral panic" as much as it being a major obstacle when trying to teach more real-world programming and data science — e.g. how to write and manage code (and its dependencies) outside of a prepackaged CoLab environment.

You really think there's no concern if data scientists of the next generation have little care for the difference between a data file read from a local path vs. a cloud path? I mean, the "reproducibility crisis" was already full blown based on old-fashioned concepts of data files...

The average user doesn’t understand the basics of an operating system, we teach that to people pursuing IT careers. If technology has changed such that what we consider “the fundamentals” have shifted, then we should be incorporating that into our education of IT professionals rather than assuming society has an obligation to make these ideas known.

Yes computer are a huge part of our life. So is plumbing, and I know very few people who are comfortable doing their own toilet installations.

But isn't the issue that it is people going into computer programming who are struggling with the concept - not just casual users, for whom it really doesn't matter that much if they don't know how to navigate a folder structure.

To extend the plumbing metaphor: it's as if no one on the plumbing course got the concept that taps had to be attached to water pipes and then wondered why they didn't work.

Unfortunately there is a fundamental difference: that of a O(1) vs O(n) lookup. While this makes little difference to the casual user, it definitely makes a difference to the power user, let alone the programmer

Reply on news.ycombinator.com