Live data from Hacker News

Ask HN: Software you hate but can't replace?

news.ycombinator.com

321–328 of 328 posts

Re: Ask HN: Software you hate but can't replace?

#321
post #165

Earlier quoted context omitted.

We're having this struggle right now trying to replace Confluence thanks to Atlassian's "no more Server licenses" foot-gun. The other wiki software that's out there just hasn't impressed me so far. This one is painful to set up and keep running, that one is awful to edit anything in (especially if you're non-technical), most of them don't support collaborative editing or inclusion of Office doc content on pages, and…

The fact that collaborative document preparation is as troublesome as this in 2022 is a damning indictment on the computer industry. (It's become very good at advertising though.) Douglas Englebart must be rolling in his grave at a comment like this https://www.youtube.com/watch?v=yJDv-zdhzMY

It really is puzzling to me that there aren't more and better options in this space. On the one side of the fence you have things like SharePoint or some of the Google bits that are fine for synchronous collaboration on a page or document through a browser, but not very tech or API-friendly. On the other side of the fence you have things like Mediawiki and Dokuwiki that are super tech-friendly, but very manager-unfriendly and not particularly good for synchronous collaboration either.

It surprises me that Confluence continues to be the only thing sitting in the middle of those, and while some competitors have started to shyly emerge (Bookstack, xWiki, I think there's one in progress at Jetbrains), nothing feels like it's really aggressively going after the Confluence market-space directly. Given how many people I've seen comment that they're not fond of Confluence but can't seem to replace it (at least, not until Atlassian forced their hand with the server-license fiasco), it seems like a natural space for people to pursue.

Re: Ask HN: Software you hate but can't replace?

#322

Earlier quoted context omitted.

I think for me, someone who is actively diving deep into Discourse and hoping it'll work because I love the search and the open-source aspect of it, I still find the UX to be off. I guess I'm used to shorter interactions online, in comments sections or on social media, and Discourse doesn't seem to do those that well. Now, that being said, they have an alpha version of chat, which is the main reason I've jumped more…

Try the Fig app. https://apps.apple.com/us/app/fig-for-discourse/id1485491193 I also recommend learning keyboard shortcuts for Discourse, especially j/k/o navigation and g-u g-n m-t m-r; it runs over most of the friction.

I like the Fig app but from what I know it's no longer being updated, and with the new chat function coming, I'd hope it would get updated for that.

I think there's just a much higher learning curve. I keep learning keyboard shortcuts but I don't know how many of my community members will. Who knows, maybe with a focus on chat will bring a new focus to mobile (and with iOS getting better support for PWAs)

Re: Ask HN: Software you hate but can't replace?

#323

Earlier quoted context omitted.

While I detest Jira, I find Confluence to be quite good for writing documentation. It's full featured and mostly just works. For our customer facing documentation at work we use Zendesk and while it's good it's nowhere near as good as Confluence.

We're having this struggle right now trying to replace Confluence thanks to Atlassian's "no more Server licenses" foot-gun. The other wiki software that's out there just hasn't impressed me so far. This one is painful to set up and keep running, that one is awful to edit anything in (especially if you're non-technical), most of them don't support collaborative editing or inclusion of Office doc content on pages, and…

> I keep hoping to find something under a rock that really wows me

Send me an email at Paul at dokkument.com and that's exactly what we are working on. Yeah unforturnely you will not see anything interesting on our landing page

Re: Ask HN: Software you hate but can't replace?

#324

Earlier quoted context omitted.

It is a horrible language, I think some called it a write-only language. But what it can do is powerful. For something there’s no alternative. They have provided a Python interface and I always want to use the Python interface more to work around the language problem. Also Sympy has a way to read Mathematica expresssion I think.

> It is a horrible language, I think some called it a write-only language. TBH, I don't think that the language is badly designed for a CAS. Most CAS (e.g., Maxima, Maple, etc.) all use a language that's heavily influenced by functional programming languages. But yeah, Mathematica is terrible for numerical work, and doesn't play well with other external libraries or tools.

I was referring to https://en.wikipedia.org/wiki/Write-only_language and seems like Mathematica is the only CAS on the list.

Actually as a functional language it is not bad. It is not the programming paradigm that it is at fault. I remember someone praised Mathematica as the best language out there because of its functional paradigm.

I’m not sure if anyone has called it a read only language too. But in some sense it is. You can look at a notebook and have no way you can write a code like that without working it through interactively. I once got one that basically there’re multi solutions and the next step chose one of the solution, where somehow which solution you pick is deterministic but you wouldn’t know without running it interactively. (There are perhaps better way to do it programmatically but the “best” practice wasn’t well known or encouraged.)

Another way to think about it is, it is not a language that one can just write a program on its own. It almost must be developed interactively. And because of that nature, it is easy to lose track when it is getting long.

Some of this are more general problem, like Jupyter which starts as something imitating the Mathematica notebook experience. But in those cases, it is much more easy to refactor things in a module and reuse them. (Where lack of namespace separation in Mathematica is a big contributing factor.)

I’ve developed a tiny Mathematica package for personal use, so I’m not an experienced Mathematica-ian, but I think what I said is well known problems of it.

Re: Ask HN: Software you hate but can't replace?

#325
post #61
post #25

Jira, for sure, and it seems hard to run away from, almost everyone use it It seems like the only alternative is to fund my own startup, with me as a sole shareholder, just to forbid Jira forever

I know Jira is difficult to get rid of once it is entrenched, but I’ve found Shortcut to be a great alternative. https://shortcut.com

Shortcut was formerly called Clubhouse for anyone who used it before September 2021 [1].

Somewhat ironically, they renamed the company to de-conflict with the other Clubhouse (the audio chatrooms app) although at this point said app has all but faded into obscurity less than 6 months later.

With respect to the Shortcut product itself, my experience with it ~1.5 years ago was positive and I would try it again.

We switched off to Jira + Confluence for feature parity that Clubhouse lacked, at least at the time. But it looks like they are very rapidly iterating on the product and adding new features (threaded comments look awesome). I look forward to giving it another shot and watching as it evolves.

It looks like they also now have a Confluence competitor [2], currently in private beta.

[1]: https://shortcut.com/blog/clubhouse-changing-our-name-to-sho...

[2]: https://shortcut.com/write-beta

Re: Ask HN: Software you hate but can't replace?

#326
post #57

Earlier quoted context omitted.

As much as they try to hide the fact, it is actually possible to join a Zoom meeting in your browser. That way you don't have to install a sketchy app with root privileges and directory traversal bugs.

I initially tried this when my language learning course had to go remote and the teacher insisted on using Zoom. But I ran into a dealbreaker issue with the web view: It does not have the grid view of all participants' cameras, and (to my knowledge) it also does not allow selecting the camera of a single person manually. This is an issue when I need to see a flash card that the teacher is holding up, but someone else…

Confusingly, there are 2 different versions of the Zoom web app:

1. The "Join from Your Browser" links take you to the worse version, which I assume is legacy or something.

2. There is also a progressive web app version [1] that has more feature parity to the desktop app. It has the grid view built-in, but for whatever reason it doesn't seem to show as many people at a time vs the desktop app grid view.

[1]: https://pwa.zoom.us/wc

Re: Ask HN: Software you hate but can't replace?

#327

Earlier quoted context omitted.

> It is a horrible language, I think some called it a write-only language. TBH, I don't think that the language is badly designed for a CAS. Most CAS (e.g., Maxima, Maple, etc.) all use a language that's heavily influenced by functional programming languages. But yeah, Mathematica is terrible for numerical work, and doesn't play well with other external libraries or tools.

I was referring to https://en.wikipedia.org/wiki/Write-only_language and seems like Mathematica is the only CAS on the list. Actually as a functional language it is not bad. It is not the programming paradigm that it is at fault. I remember someone praised Mathematica as the best language out there because of its functional paradigm. I’m not sure if anyone has called it a read only language too. But in some sense it…

> I was referring to https://en.wikipedia.org/wiki/Write-only_language and seems like Mathematica is the only CAS on the list.

Interesting, and thanks for the link. I didn't know such a term existed, although I'm a bit disappointed to see C and Lisp in the list, which some consider to be extremely well-designed languages.

> Another way to think about it is, it is not a language that one can just write a program on its own. It almost must be developed interactively. And because of that nature, it is easy to lose track when it is getting long.

I agree. Although Wolfram Inc. has been pushing WolframScript and Wolfram Engine (and the IntelliJ IDEA plugin), at the end of the day the Wolfram language is best suited for interactive development, i.e., through the Notebook interface. And Mathematica Notebooks do not play well with most version control systems and are best treated as binary files.

Re: Ask HN: Software you hate but can't replace?

#328

Earlier quoted context omitted.

I was referring to https://en.wikipedia.org/wiki/Write-only_language and seems like Mathematica is the only CAS on the list. Actually as a functional language it is not bad. It is not the programming paradigm that it is at fault. I remember someone praised Mathematica as the best language out there because of its functional paradigm. I’m not sure if anyone has called it a read only language too. But in some sense it…

> I was referring to https://en.wikipedia.org/wiki/Write-only_language and seems like Mathematica is the only CAS on the list. Interesting, and thanks for the link. I didn't know such a term existed, although I'm a bit disappointed to see C and Lisp in the list, which some consider to be extremely well-designed languages. > Another way to think about it is, it is not a language that one can just write a program on it…

I agree C kind of stand out from that list (I don't know lisp). But I think as C is a small language, some people can write pretty simple to read C. But I guess the term is used to describe C as well because of how "close to metal" it is, and people can do some insane things to optimize their code, in a way that becomes "write-only". One example might be Linux kernel code?

Another example from what I heard directly:

A few years ago, from one of the project nominated for the Gordon Bell prize, one of the Scientist showed us their code. For perspective, this was to be run on Intel KNL so it is extra hard for the compiler to optimize it well. Since our machine is from Intel (and the contract stated they will provide support), Intel engineers were involved in optimizing a kernel. In the end they generated hundreds of lines of C code and committed to the code base directly. It was partly because it is time-sensitive so they can't wait until they modify the Intel compiler to optimize that code, so they sort of optimize it manually. Now if you read that source code, there's no way you can modify it, rightly fit into the description of "write-only" language.

Also, there's a less well known term of read-only language. I can't find a wiki page for that. But I think AppleScript is on the list, because you know, reading an AppleScript makes it looks quite straight forward it should be written this way, but if you're asked to reproduce it or write something similar, you can't. Hence read-only.

Edit: It was described in earlier history of the same wikipedia page but was removed: https://en.wikipedia.org/w/index.php?title=Write-only_langua...

Post reply on HN