Hacker Newsnew | past | comments | ask | show | jobs | submit | spv's commentslogin

SumatraPDF is great! It’s fast, but does not work well for annotations like highlighting text.

I have switched to Okular, you can install it directly from the Microsoft store directly.


As some one who reads pdfs all days with lots of math in them, I cant go live without my 27inch 5k display. This is post is on point.


Please add math support in github markdown. Something real users need.

Stop trying to make it a social network.


We're definitely going to add math support to GitHub markdown.


Is there a timeline for adding math support to GH markdown?

I've had to resort to human-side rendering to put LaTeX/math in a GitHub flavored markdown file.


github being a social network is the main reason of its success.


Is it? Networking yes, social maybe?

I don't have any memories of early GitHub but did any of it's competitors come close to being as good?


I'd rather that they supported a standard like MathML or LaTeX.


There was someone working on LaTeX support a few years ago, but iirc it was very difficult to secure it enough to feel comfortable letting it run server-side for rendering. Source: I worked there when they were working on it.


What about client side rendering via MathJax? You could have your LaTeX code embedded in your markdown and MathJax could substitute it for appropriately rendered symbols as necessary.


It's kind of hard to imagine how DoorDash execs are going to explain to those investors who find this on HN / elsewhere and ask them what the eff is going on ?


Can nbdev be used with other Jupyter kernels, for eg like Julia.

My main issue with notebooks is putting them under version control in git. This seems like it could help with that.


Yes, nbdev can help with merge conflicts for notebooks [1]. Also checkout,

- GitPlus[2] - JupyterLab extension for git version control

- ReviewNB[3] - For notebooks diffs & commenting

Disclaimer: I built GitPlus & ReviewNB

[1] https://nbdev.fast.ai/#Avoiding-and-handling-git-conflicts

[2] https://github.com/ReviewNB/jupyterlab-gitplus/

[3] https://www.reviewnb.com/


reviewnb looks interesting. cool project.


I use NBDime to make readable diffs of Jupyter notebooks https://nbdime.readthedocs.io/en/latest/


this is exactly what I was looking for. thanks.


Jupytext (https://github.com/mwouts/jupytext) solves most of the problems with managing notebooks in git.


nbdev works with any kernel, because it only cares about the notebook files, which as you know are just JSON.

I've been using it with xeus-python kernel lately.

Edit: Forgot the Git part. It was designed to handle the git limitations in notebooks as well. In fact, one of the commands in the tutorials is to set up it's built-in git hooks, which will run special clean commands on commits.


I cannot believe that you make more from such a niche dev tool than your job at Google. I always thought people who use clojure/scheme would be using their custom setup in emacs or vim.

I'm interested to learn more details, how things were when you first started out selling the app and the trend.


You say that but one thing that i always dislike when i consider taking a look at Lisps is how everything seems to be either on Emacs or looks like it'd really like to be Emacs.

Personally I want a full blown IDE that takes advantage of advanced modern technologies such as displays that can draw individual pixels, have a model of the codebase that allows advanced features such as word completion and preferably fits nicely with the underlying OS. A debugger would also be nice, but i understand that sometimes i ask too much.

I wouldn't mind paying for such a tool (though i do mind DRM schemes and subscriptions - i want to be able to pay once and then be on my way). Cursive looks something i'd pay for if i was really interested into Clojure and was using macOS.


Actually, there's completion, and excellent debugging tools for Clojure in Emacs, both step-by-step kind, and investigative. "I don't want to use Emacs" is a completely valid stance (de gustibus...), but it should not be misinformed.


I know that Emacs has good support for Lisps in general, my comment was a joke towards what i consider basic features of an IDE (the whole 'modern' and 'advanced' thing when referring to stuff that were available in the late 80s should have made it obvious :-P). It comes mainly from my observation of the trend where a lot of developers like to jury rig "IDEs" out of text editors and a bunch of other unrelated tools that look as if they are made to be run in 70s terminals that look and feel considerably worse that MS-DOS applications which had 1/100000th of the available resources.


> Cursive looks something i'd pay for if i was really interested into Clojure and was using macOS.

um... it runs on Linux and Windows as well.


Ah, somehow i missed that it is build on IntelliJ. I thought it was a native macOS application.


FWIW I've found Clojure has pretty decent support in a few editors. I know Cursive is quite popular, but I've also used both VS Code and Atom to some extent, and they have nice features.

Somebody else could probably speak more to it than me, since I do primarily use Emacs, but I definitely don't introduce my friends to Clojure with it.


> I cannot believe that you make more from such a niche dev tool than your job at Google.

It's important to note, that there are markets where Google pays 70-90k / year[1].

[1] https://www.glassdoor.com/Salary/Google-Berlin-Salaries-EI_I...


I'm not talking about those markets - this was after an acquisition and I was supposed to move to either SV or NYC. In the end I didn't want to work at Google or move to the US so it was pretty brief.

Thinking about it, I'm not sure Cursive pays more than my total comp at Google, but certainly more than salary + bonus.


There are markets where it pays less - see Warsaw.


I wrote a little bit about this below (https://news.ycombinator.com/item?id=21337892), but I think that a lot of that perception is because lisps (with the exception of CL) have traditionally been things that people tinker with as a hobby. Once you get a significant number of people making a living working with a language, suddenly a lot of people want barriers to getting their work done removed. And Emacs, for all its benefits, can be a significant barrier to getting work done for a lot of people. Plus there are those of us who just think that life is too short :-)


As a Clojure (and Java) dev, I started with Cursive and eventually moved to emacs w/CIDER as I became more comfortable with it. Cursive is great for people new to the language who may not know emacs - learning a new language is challenging enough without also having to constantly refer to an emacs cheat sheet :).


FWIW Cursive is also great for people who have a lot of experience with Emacs and just can't be bothered with it. Popular examples include David Nolen and Timothy Baldridge (and me), but there are plenty more too.

One representative comment that someone made to me at a conference was along the lines of "Since my boss encouraged us to move to Emacs, I've never spent so long fiddling with and arguing about my editor config. After a while most of us moved to Cursive and all that just went away".

I think Emacs is generally easier to get along with these days, but it'll never be as easy to use as a "normal" app where things just work out of the box. Some people like that, and others don't, but it's not as simple as a division between newbies and experienced users. They're roughly equivalent in power these days, some things Emacs does better and others Cursive does better - fortunately there's plenty of room for both, even in a userbase as small as Clojure's.


I think if stackoverflow conducted their survey on HN readers alone, the editor-used distribution would be very different.


A lot of people use Emacs, but I remember as Cursive took off in the Clojure community. It's so wonderful being able to work on a large JVM app that includes Scala and Java in IntelliJ and have first class Clojure features. Cursive is phenomenal, and because it's (or was at the time) the only choice, I would gladly fork over 2x the cost out of my own pocket (though my company would let me expense). It was great too to go to Conj and major Clojure conferences and hear Colin announce new features and talk about the roadmap.


My team is mixed. We have two devs using Cursive, one using Atom, one using vim and then two using emacs. The two devs using Cursive love IntelliJ and being able to work in IntelliJ and do Clojure development is a big win for them


They (we) used to, but with Cursive.... welcome to the XXI century :)


I have been using Bitwarden for some time now. It’s an open source password manager. There are apps for all major platforms and extensions to all major browsers. Checkout https://bitwarden.com/


Interesting in that you can host your own instance of their cloud server, but I really prefer something that uses standard cloud storage mechanisms (Dropbox, iCloud, etc.) for sync and works on top of that.

Like others here, I'll probably be reevaluating my choice when it feels like it's time to upgrade. For me, some of the open source solutions are perfect as far as the underlying storage format and sync technology, but lack good browser extensions that already understand all the quirks of various sites. That's the kind of thing that a commercial product can tend to do a better job at.


I use Enpass for this, which uses any kind of regular cloud storage backend and has a fully-featured desktop client and browser fill plugins for free. The mobile clients cost money ($10 per platform, once) which I think makes perfect sense.


It's simple enough to do with some version of KeePass. I'm currently using https://keeweb.info/ and Keepass2Android.


I use the same. The only PITA is unlocking it with my, rather long, pass phrase. I just find it really fiddly on mobile.


You can get it to recognize fingerprints or a short version of your password if it's even been fully unlocked for the current phone session. It's a little fiddly and may not meet the level of security you're looking for, but it's an option.


I knew about the short password, but hadn't seen the fingerprint option - that's just what I need to make it "secure enough" while being usable!


What do you use to fill browser login/passwords? There seem to be so many plugins for keepass


I don't use any plugins. I just copy/paste, which if you're doing it from the app stores the copied parameter in memory for a ~15 seconds, after which it is flushed.


1password is also supported by many mobile apps (on iOS at least).


I've been an early adopter and really they are coming in leaps and bounds. The only complaint I have is that integration with iOS apps is very very spotty, but I believe that's an issue with Apple muscling 3rd-parties away from that field (and to a certain degree, an issue with developers not following best practices in their apps).


+1 for bitwarden. Recently I've become a subscriber because I love their service and I wanted to support them.


Looks awesome. The only drawback i see is it doesn't support iCloud.

Personally i wouldn't want to store my passwords with the same company that creates the (encryption)software.

And i definitely wouldn't self-host (why burden yourself with the management of data-backups/updates/etc. when iCloud is available for free...)


To be precise, their base software is Free software, licensed under the AGPLv3. The also distribute non-Free (and non-Open Source, and non-gratis) software.

Their base software has an artificial limit in terms of number of users and number of 'collections', which goes contrary to the ethics of Free software.


This is not correct. All of Bitwarden source code is 100% open source. Even the few features that are paid. https://github.com/bitwarden


How can this be an issue if their base software is open source? Just recompile it with the numbers upped, surely?


Gah, you're right. Their paid services and limits seem to be related to their SAS offerings.


Your passwords are stored on their server. You'd have to compile and run your own server, which is more expensive than the $1/month they're asking for.


So you’re paying for the service they offer: a hosted version. You do so because it’s cheaper than hosting your own. There’s no conflict at all with any open source ethic.


Yes, that's exactly what I was saying.


> You'd have to compile and run your own server, which is more expensive than the $1/month they're asking for.

For people like me that already rent a VPS for their mail and website the marginal cost is $0 except for the time it would take for me to perform the installation and setup.

If the system is good and stable then the "cost" of the time that I would spend installing it on my server would be close to $0 when divided over the amount of time I use the software in the future.


I think another plus of buying their service is your supporting development of the software and saving yourself time, while a critical piece of your security software remains open source.


> which goes contrary to the ethics of Free software.

No it doesn’t. Free software doesn’t have to be free: Even on the GPL page it’s written that it’s even ok to sell free software. It’s only unethical if you equate OSS to software communism, but that’s another topic.

So anything that encourages the user to either use the freemium, then either dive into the code or either pay, is ethically correct. After all, you can download their AGPL, knock the limit, and redistribute. At which point you’ll be a contributor and while you’re at it, you’ll probably make a few other improvements: it means effectively free for contributors, which is awesome. See, it articulates quite well gratis, contributors and funding.

It’s only designed to make enterprises pay, which is good because they can “donate” huge sums for good software, so it funds the open-source community quite well. And it retains the qualities of OSS: You know what you install, you’re not tied to the editor if he dies, and if they stop improving the software, a contributor can take over their code and become more famous. Win-win-win.


Boo, I was looking for something that had a cli and integrated with every browser as well as had an ios app.

I've not looked again recently but lastpass is the only thing I've found that fits those bills.



I saw all that but it looks like you've got to pay a monthly fee for full access for their cloud storage. Not sure its worth the hassle of migrating to in this case.


Asked my buddy what he was using and he pointed me to this: https://www.enpass.io

Seems to fit the bill a bit better. And might mean $12/year less though thats not a huge problem.


Maybe, but Bitwarden is open source, and Enpass is not. It's not important for people unless it is important for them, and in that case it's usually very important.


It's $1/month to support open source software. That's not a bad price. You can also choose to use their hosting for free too; I was doing that for a few months before fully adopting it.


I already pay lastpass $12/year. Not sure its worth the hassle to convert all my login information.


For me the default colors on ubuntu were never welcoming. Changing them would be great. Hopefully with Unity 8 they choose better and warmer colors


I don't understand why Gadfly is the top repo. I always found it to be very slow and having too many dependencies. Last I remember it installs more than 20 packages as dependencies. Its also not a supported backend in Plots.jl


Gadfly was one of the first Julia packages, and the first plotting package. That is probably why.


Wire also takes a similar approach.

I think google is trying to mimic whatsapp here: A single device with end to end encryption so that your keys never leave the device.


Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: