Hacker News Reader

Comment by Mediterraneo10 | Hacker News Reader

There have been complaints about the UI since forever, and the developer is well-known for bristling at any constructive criticism and telling people to either accept the way he likes things, or simply don't use the application. Consequently, I am surprised that no one has forked Calibre, even if it was just to kept the backend but simply overhaul the UI.

Something that I would like to see in a future version of Calibre (or a fork) is news-download recipes changed to a plugin system that can be updated separately of Calibre itself. It is not ideal that if a news recipe becomes obsolete due to changes on the respective news website, one has to upgrade Calibre to a new version. For users whose distros package only a certain version of Calibre, it would be nice to continue running that distro-supported version but simply be able to update the news recipes from within the Calibre settings.

View on HN

The code base is a nightmare and still uses Python 2. I'm not surprised it hasn't been forked, although I'm a little surprised it hasn't been replaced.

Frankly, the software is just fine for most needs, and the person in charge is still active, so the desire to replace it just hasn't materialized.

EDIT: Looks like somebody actually took the initiative to port it to Python 3.

Calling the codebase a nightmare is the understatement of the year. I tried forking it at some point. I don't think I could reasonably even attempt describe the codebase in a way that wouldn't get me banned from HN, and most likely the rest of the internet.

Needless to say I gave up.

We would actually love to read that.

An article title such as "Popular open source software Calibre has a codebase that isn't just bad, it's outright dangerous" would definitely make me click.

If you feel strongly, please do this. It adds value. The creator might really benefit from it.

How would you feel if someone came out with an article about "echelon's code isn't just bad, it's outright dangerous" and wrote a whole article attacking and shaming you, just because you might benefit from it?

Sure you might benefit, but there are much more constructive, kinder ways to do it. Even if he and his code aren't great, Goyal's done a lot of free public service here. Let's not be dicks.

If you want to learn about the codebase, here it is https://github.com/kovidgoyal/calibre

Good point. I was responding to the tone of the OP, but it's not at all kind or constructive.

They should be praised for their contributions, not torn down. That would be a dick move.

Forgive me, I have the flu and I'm grumpy today.

No. I definitely wouldn't like to read a post with an attitude like the one espoused in the titled you proposed.

However, I would gladly read a post that reviews Calibre codebase in a helpful, friendly way.

> If you feel strongly, please do this. It adds value. The creator might really benefit from it.

There is an awful lot of self-deception packed into this paragraph.

> We would actually love to read that.

That part is probably true for many.

That’s clickbait and not very constructive. Fun to read, maybe, but it probably won’t be for Kovid and it won’t help make Calibre better.

That makes me very curious. Could you write a blogpost on the topic by any chance?

> I'm not surprised it hasn't been forked, although I'm a little surprised it hasn't been replaced.

Shouldn't be all that surprising. Things are harder than people think, and only a few people are willing to stake actual skin in the game instead of just complain.

Also, Calibre is already a niche tool. All of my friends/family have a Kindle, they all use Amazon's default send-the-kindle feature, and none of them probably even know how to get the physical ebook file out of the Amazon system nor care.

And ebook file management just isn't a very sexy problem to solve beyond a proof of concept weekendware effort. Surely most people aren't spending much time in Calibre anyways. I know I just use it intermittently to load the next 10 books onto my ereader. It's not a part of my life I'm very passionate about changing.

> EDIT: Looks like somebody actually took the initiative to port it to Python 3.

For reference, here's the ongoing python 3 port pull request: https://github.com/kovidgoyal/calibre/pull/870

I would say that the reason it has not been successfully forked is that reliability is better than a nice UI.

It's an incredibly reliable piece of software. Isn't that what's most important?

What's so problematic about the UI?

It opens with a list of books, you double click on the book you want to read and it opens that book in a new window.

What could be simpler than that?

What are the complains? Just that it looks like an old 90s program, and doesn't have the latest all-white minimalist design? Or are there actual UX programs?

It looks like an old 90s Windows program, and most Mac users were bothered by those even at the time. It fails to honor basic common MacOS navigation keyboard controls, making it jarring even if I tune out the over-the-top ugly visual interface.

Old 90s Windows programs are great. Downright relaxing after the "WTF is interactive and what's content?" mess of "modern" UI. Calibre's not that. Some poster upthread got it right by comparing Calibre to some of the uglier examples of Java UIs.

(mind, I find Calibre pretty usable)

The code is a very ugly mess.

>It is not ideal that if a news recipe becomes obsolete due to changes on the respective news website, one has to upgrade Calibre to a new version.

I haven't used the news feature in almost a decade, but in those days, you could clone a news site's recipe and modify it - no need to upgrade Calibre.

Recipes are first polled from a central repo so they do update if the app is online.

Reply on news.ycombinator.com