I'm looking for feedback on my distributed log db project. It's a part I pulled out of a stochastic time tracking system that we are building. I think it's a neat idea and has a lot of potential and I haven't found anything like it yet.
Hi HN! I've been working with the Stellar blockchain recently and I found it hard to get a simple, open and transparent wallet that would let me keep my accounts on my local machine. So I wrote my own and made it available for everyone.
I've tried to keep the code simple and clean and I would appreciate any feedback. The goal is it should be a simple, powerful wallet that any dev can pick up - be confident how it works and be able to hack in 5 minutes.
28 years! It frightens me and boggles my mind how something like this can happen after being married for so long. Looking back do you have any idea why? I'm sorry for your loss and I hope it'll get better for you soon. As others have said it's perfectly normal to feel depressed at such a life altering event and it probably does take a year or two to recover.
Some chronic depression before, maybe. But mostly, she said she wanted to find herself, and she couldn't do that with others around -- too hard to "hear herself think".
Which is to say, I don't really know and never will.
One trick I've learned is to write out, (in a structured comment), what I am going to do an how I'm going to do it _before_ actually writing the code.
It seems redundant but it actually forces me to do a quick think-through of what's happening. Plus distractions get a lot easier to handle: I just need to read what I've written down to reload the context.
Paradoxically, I've found that this method of typing out _more stuff_ actually helps me finish much faster.
I second this. Over time and struggling with concentration in bullshit jobs, I've learned following tricks when I feel stuck and/or stressed out:
- I have an org-mode document with a TODO list and project notes always open. Whenever I'm stuck, I'll start decomposing the task into smaller and smaller TODO items, simultaneously with doing a written brainstorm. Sometimes I'll write out a hundred lines of text, but it does help me get unstuck.
- Similar to what you mention, I frequently write a TODO list in comment as a scaffolding in the file I work on, and I then proceed to fill the space between TODO items with appropriate code.
(off-topic) Org mode is really amazing. I've transitioned almost completely from Emacs to Vim over the last few years but I still have Emacs on my dock only because of `org-mode`. I've tried substitutes in vim but nothing really works.
(back on topic) Another advantage of writing comments is coming back to the code becomes so much easier. I can re-read the comments and understand what's going on much faster.
To do this properly I've developed a simple system of 'categories' that allow me to structure the comments properly. I think they works really well. For anyone curious - here's an example in my last weekend project: https://github.com/theproductiveprogrammer/luminate
Check out the comments in the code (`main.js`). I think they make the code much easier to understand and maintain. It's almost an alternative to TDD - Comment Driven Design (CDD!)
Hi HN! I've been working with the Stellar blockchain recently and I found it hard to get a simple, open and transparent wallet that would let me keep my accounts on my local machine. So I wrote my own and made it available for everyone.
I've tried to keep the code simple and clean and I would appreciate any feedback. The goal is it should be a simple, powerful wallet that any dev can pick up - be confident how it works and be able to hack in 5 minutes.
I don't know Stellar, but I am interested in crypto and blockchain technology.
You say you are working with the Stellar blockchain. But then you say it's a thin wrapper around an API.
Are you doing the crypto (creating private keys, signing transactions) in your code? Looking through the code it looks like you call something called 'stellar-sdk' for it. Which is not part of the repo. I have the feeling the magic is happening in this external code. Which conflicts with the statement that the wallet is "easy to understand at a glance and transparent enough to be easily customized and fully trusted". One would have to also analyze that stellar-sdk: https://github.com/stellar/js-stellar-sdk/tree/master/src
Something that is way too big to understand easily.
Also, I see some account creation code in your repo. If it is using a blockchain, why is it creating accounts on a service via an API? Can't I have my Stellar assets under control without having an account somewhere?
They provide a REST API for interacting with the network itself and JS/Go/Python/Java standard libraries that can be used. Of these, the JS library is the most developed which is why I picked it.
Using the standard stellar library to connect with stellar is secure and transparent enough. What the wallet does is encrypts and protects your account keys (which are what give you control over your accounts). It does this using `scrypt`(https://www.npmjs.com/package/scrypt) for PKDF and sodium (https://www.npmjs.com/package/tweetnacl) for the actual encryption. Having access to the code should show you that there are no backdoors built in, no information being sent across to any third-parties, and so on. This is what people care about when protecting their accounts.
Your second question is also very insightful. Creating accounts is at the heart of Stellar. You cannot start performing any transactions on the network without first creating an account. And, in order to create the account, it needs to be be funded. This is why the wallet allows you to create (and fund) a new account on the network. You can read more about it here: https://www.stellar.org/developers/guides/get-started/create...
There could be bugs, of course, but the library is open source and developed by the Stellar network team.
Of course you are right in a sense - the core team also should not be blindly trusted. But, because your account is managed by the wallet and stays local to your machine, there is nothing much the library _can_ do IMHO.
It's code that runs on your machine. What would prevent it to do whatever it wants? Like sending your private keys somewhere? Or using a weak form of crypto to create them so that somebody can guess them?
That's an excellent point. The only place the library has access to the private keys are during account creation and message signing. Perhaps I should implement both directly in the wallet so it's visible and clear to everyone.
Nim is a great little language - it is enjoyable to work in and small enough to learn easily. It's neat for little utility programs that are "too big for bash" and "too small for java/C". Sometimes I think of it as an alternative to python.
I have so many things that I need a system to keep me productive. Plus, it's fun to get things done and be confident in not having missed anything important.
Productivity is about getting the right results for the least effort!
I found this advice on negotiating to be most helpful:
1. Always stay polite. Getting emotional works against you in most cases.
2. Don't give ultimatums. Saying "I want this and this" is fine - don't add "or else...".
3. In your current situation, you have lost some negotiating leverage (your previous salary, your experience) but it's not unsalvageable. Simply bring them up and counter them yourself:
* Your previous salary is not relevant - only what you want to accept now. HR has a desire to lowball you so you should counter with a higher-than-market request.
* Your experience is not an issue. If they want you for the job then you are qualified for the salary that job entails.
4. Use the fact that you are _already_ giving them something by accepting an offer one level below! Make this a big deal and ask for compensation for doing a higher level job.
5. Pick a price that is 10% above market value and have a list of reasons you want it (new house, car, ...). Do not be afraid to stick to your ask. After all, they need you to do the job just as much as you need their job.
6. When picking a number do not pick a round number (numbers like 30% sound like you have not given it much thought. Numbers like 101,300/- sound like you have figured out your finances and have a good reason for your ask).
7. Be patient and remember you both want something out of this.
Good luck!
Edit: The Lightning Guide to Software Resumes has a short section on negotiating that is a good summary. A better book is "Never split the difference" by Chris Voss which is, I think, the best book on the subject.
I'm looking for feedback on my distributed log db project. It's a part I pulled out of a stochastic time tracking system that we are building. I think it's a neat idea and has a lot of potential and I haven't found anything like it yet.