[hereonout2]

There was a lot of quirkiness and egotism then, the cult of personality was rife, not just with project leaders such as DHH but ancillary dramas from folks like Zed Shaw.

The stuff from _why was always wacky but some of their projects were also really inspiring for younger programmers, of which I was one. I could take or leave the cartoon foxes but I remember seeing the code for the Camping framework when it was released and being amazed at how simple and elegant it was.

The ruby and rails community set a path that is followed by many to this day. We're all python programmers now but there'd be no Flask without Sinatra, or Django without Rails. The author of the excellent dependency management tool Bundler went onto work on Yarn for javascript and then created Cargo for rust. Mitchell Hashimoto wrote Vagrant in ruby before starting hashicorp, there's loads of examples like this

I disagree that the Poingant Guide was the primary documentation for anything, it was a quirky guide for new comers. The Pragmatic bookshelf had the definitive guide to the language, as well as the rails book and then a whole series of other publications. I also remember spending a lot of time on the core ruby documentation, which does a fantastic job of covering the extensive standard library.

[throwanem]

Ruby itself was tolerably well documented, but that was a rare bright spot, and of limited utility when Rails leveraged the malleability of the language into near incomprehensibility even by the language's own low standard. And while you're not wrong that by spending enough money it was possible to obtain halfway decent documentation for Rails that if you were very lucky might still be mostly applicable after a couple of point releases, not everyone had a lot of money to spend in that way. Pay-to-play access to knowledge required for basic competence is another example of structural exclusivity, and one that stood in stark contrast to most contemporaneous projects of similarly high profile, by comparison with which Rails had all the openness of a Freemason lodge in 18th-century Italy - oh, they advertised themselves very effectively, that I grant you, but in terms of actually fostering development among those so attracted, I think a secret society in which membership could be a death sentence would probably do a lot better.

> The ruby and rails community set a path that is followed by many to this day.

I've already acknowledged this, albeit without the unwarranted gloss. I might instead have said that Rails' one unqualified success was as a fount of good ideas badly implemented. But since on reflection I'm pretty sure Rails' use of ActiveRecord popularized ORMs in web application development, I suppose it isn't even fair to say that all the ideas were good.

[Tainnor]

Hibernate predates ActiveRecord, if I'm not mistaken, and to this date I can't decide which of the two I hate more. They both do different things very wrong. So I'm not sure if you can blame ORMs on Rails.

Also I'm not sure what you mean by the documentation criticism. I criticised Rails elsewhere in the comments, but I can't really fault it for the documentation, which IMHO was always excellent. And then you also had Michael Hartl's excellent Rails guide (which was available online freely, at least at the time) which was what basically taught me modern backend development (including what automated testing is).

[throwanem]

I'd have loved to know about such a thing at the time! One wonders why none of the people in the various IRC channels and forums where I then sought Rails advice saw fit to mention it.

In entire, albeit mildly grudging, fairness, I do have to concede that Rails introduced me to the concept of unit testing. But I'm still glad I learned modern backend development in the years immediately after Rails peaked, and while the worthwhile was being sorted from the nonsense among the many concepts and approaches that Rails does, for better or worse, deserve credit for having made newly popular.

[hereonout2]

But rails was modern backend development back then, it offered an alternative to a world full of mod_perl, PHP4, Java servlets with JSP and XSLT.

Having had the displeasure of working on all those stacks it's hard to overstate how transformative rails was at the time for me.

Built in unit testing is one thing, not having to write a java class to expose a custom function to an xslt processor in order to format a string is something else.

[gensym]

> But since on reflection I'm pretty sure Rails' use of ActiveRecord popularized ORMs in web application development

As someone mired in enterprise web application development in the early oughts, I can tell you this is kind of backwards. ActiveRecord was actually a simplification of the more complex ORM that enterprise software (mostly in Java at the time, but a little bit of Objective C in the banking realm) was using. I _think_ the term "ActiveRecord" was first used in the 2003 book Patterns of Enterprise Architecture, and it was described as a pattern you could use when you didn't need the complexity of a full-blown ORM. For people who had wrestled with Hibernate or WebObjects, ActiveRecord felt like a light-weight sigh of relief.

That said - even as someone who still works on RoR apps - I'm glad we've mostly moved beyond ORMs (primarily by moving beyond objects, which were never a very good fit for representing data in the first place).

[throwanem]

Huh, okay, that's fair. ActiveRecord was the first ORM I worked with, and ActiveRecord in 2013 was easily poor enough to color my perspective on the category. I've heard Hibernate criticized before, of course, but I didn't realize it both predated ActiveRecord and was so much worse as to leave ActiveRecord even in its day looking good by comparison.

I still think Rails gets the blame for popularizing the concept, but I suppose that has to be mitigated by prior art making it so easy to popularize - "it's just like what you're used to, but won't make you want to kill yourself to use" is a pretty compelling pitch.

[gensym]

LOL. I feel like a geezer whenever I talk to people about what enterprise software was before RoR came along. XML. So much XML. Do you think that you should write XML in order to query data from a DB with a perfectly serviceable query language? People sure did think that in 2001!

[throwanem]

You joke, but honestly, it's still the same class of problem. I don't think I should write TypeScript in order to query a DB with a perfectly serviceable query language, either, and I've seen so much time get spent on dealing with the headaches condign upon leaky abstractions to render the productivity benefits claimed by ORM proponents transparently nonsensical. And yet...