Hacker Newsnew | past | comments | ask | show | jobs | submit | mickael-kerjean's commentslogin

In section 6: "tar files ... of the Linux kernel sources ... version ... 1.99.10 ... are approximately 24MB in size ... Out of the 2441 files in the 2.0.0 release 291 files had changed"

It never crossed my mind Linux at some point only had 2441 files and you could actually parse the code that went through a new version, that time has sailed



Whenever anyone writes about mongodb or redis I hear it in that voice.

is there some open source product who can leverage this or this just assume you have to use Microsoft stuff?


The SMB Direct support mentioned here is in the kernel for client & server from 5.15+. After that it's just a mount point any application can access. No Microsoft stuff needed on either side.


This was probably added by/for Tuxera to increase Tuxera Fusion SMB performance.


I think it got added by Samsung / someone employed by Samsung at the time https://www.phoronix.com/news/KSMBD-Lands-In-Linux-5.15


I launched Filestash [1] as my response to the infamous “Dropbox should just be FTP” comment. Once I had a decent FTP experience, I kept going: adding support for pretty much every storage protocol, plugins to expose Dropbox (or anything else) over FTP, SFTP, MCP, or S3, and all the features I wished Dropbox had, with plugins to customize everything.

The base product is open-source and I make money from custom builds, additional plugins, paid support, and the occasional extra feature for companies with specific needs. It's a bit more than noodle profitable but quite under a normal salary.

[1] https://github.com/mickael-kerjean/filestash


Where is the dropbox-part in this? This seems to be a filemanager for remote storages, which is kinda the opposite of dropbox, which is mainly a local service for syncing data. Or did the documentation missed explaining the sync-function?


What a blast from the past. I attempted to build a file-sharing tool for my team when we had video and images strewn across the org. I prototyped embedding filestash for the frontend.

It was basically a backend for generating STS credentials on the fly using a more ergonomic interface. It never went anywhere and I haven't thought about it in years, but I still believe it was a good idea that I just didn't have the organizational clout or time to push forward.

Edit: apparently I contributed at some point too? I *barely* remember that. Glad to see the project is still succeeding!

[1] https://github.com/subdavis/workspaces-io


This could probably be more profitable for you with better tagline/explanation!

I have the context for what "that comment" was, might even be in the target audience, yet from the landing page I'm still not entirely sure what it actually does. Might be worth trying a few "it's like X but with Y" or "imagine if dropbox could Z" and other formulations on uninitiated people in your target audience?


I found a typo: "Apply fined grained access control to keep your shared content under control." should be "fine-grained access control".

There: https://www.filestash.app/smb-client.html


The faux screenshot of HN tickled me.


The benefit is "Distribution". If your users are there, you want to address them wherever they already are, this is why apple store / play store / amazon store ... are so popular. Becoming a platform / ecosystem is the common playbook to go from being a one product company to an ecosystem / platform worth a lot more


Can a small business succeed in that game?


I had a X250 who died less than a month after the warranty, now got a E14 since ~2 years and it got keyboard issues


Is this some sort of supabase but with sqlite and without the operational complexity?


Yes that is what it is.


emacs is quite approachable because of its self documenting nature. In practice, if you want to drill down on a particular thing, you need to use some of the describe-* commands. For example want to know how emacs save something?

1) find the command it uses with describe-keybinding, you find the command "save-buffer"

2) "describe-command save-buffer" brings you onto the lisp world where it is defined => "files.el"

3) want to know how a variable is define within files.el? "describe-variable buffer-file-name" and now you are in C territory

4) rinse and repeat with some describe-function when needed

5) get lost onto the beauty of emacs which in my opinion is its interactive / self documenting nature which unfortunatly is not more common in all the software we use


My first job out of uni was in creating automated tests to validate some set top box. It involved using library of "blocks" to operate a remote control. Some of the people I have been working with spent their whole career in this narrow area, building those libraries of block and using them for customer and I have no doubts a LLM can today produce the same tests without any human intervention


This is not as simple as it sounds. Just yesterday I had a call with the Delft university of technology in Netherland, they want me to add some features on the free version of my FOSS product [1] but they did not want to pay anything. Over the last month, I was in contact with a 800B publicly traded company for a 1.8k per year invoice, once we agreed on the general direction they kept adding expectations, first was to sign tons of paperwork with their security checklist, legal stuff which took a few days but when they start asking for things that would take potentially weeks more, I invite them to do extras on a contracting basis, since them I have never heard back and of course they never paid a dime. I have literally tons of stories like this from governments to F500. In my bubble the paid support plan mostly work with US entities.

[1]: https://github.com/mickael-kerjean/filestash


Universities are a special case. They generally don’t spend money because of the red tape.

In much of corporate America expenses under $100 give or take don’t even require documentation, so a $50/month support subscription is easily purchased.

Just need to find the person with the purchase card.


In my experience, having MIT and UCI as customers, US universities are much easier to deal with small to no process for simple cheap things. On the other hand, I was contacted by a well known engineering school in France (ENSEEIHT), they wanted support but were laughing at the idea to spend 20$ per month for the privilege, left the impression they wanted to use my time for way under minimum wage, same yesterday with a deutch school who wanted help but not willing to spend a dime, and some other universities who have deployed my software in prod but did not upgrade in the last 5 years. Even in China, I stumbled upon a fork maintained by the university of Shangai, of course they never reached out in the first place to ask for any kind of support, just took the code and went their own way. This kind of behavior haven't happen with US universities which are more likely to reach out and pay for support


> I stumbled upon a fork maintained by the university of Shangai, of course they never reached out in the first place to ask for any kind of support, just took the code and went their own way.

I think this one is 100% ok for open source, no?


this is fine indeed if you didn't ask for my time


"Pull requests welcome"


It is very hard to find a person with that purchase card.

As a developer in corporate environment I won't get anywhere close to be able to influence anyone to buy a support, or a subscription for an open source or closed source product. This is my third corp, and it was true in all of them.

The most what I got is the approval to do some PRs for such projects during company time.


It's actually pretty simple. For the former case, you do nothing. Tell the university to find someone to make the improvement, or do it for fun. In the latter case, you should be charging 5-10x that, for starters... You send a Statement of Work, and only do what's in that, and only after they pay.


You're too cheap. Anyone that won't pay for a proper enterprise support contract you should tell to pound sand. You'll be surprised that when you start charging more people will actually take you more seriously and will be more inclined to sign up. It's counter intuitive from your side, but perception is reality. A 20k/yr enterprise support agreement is more believable to provide results than a 2k/yr deal.


Quite. I remember one of my first corporate customers who was very suspicious of $2K/week, because nobody could do that for that cheap they said. It was nothing extraordinary, just some integration work, tests etc for the project, they wanted it to work with their other suppliers' systems


I agree, it's not as simple as "$500 per year". In some cases it can be, but mostly it's not.

Firstly, you need to clearly define what is included, and even more so, what is not included. How many hours is $500? Who decides what us should bug? Can they get new features because they have support? How many installs does the support cover? And do on.

And if they start with things like "supplier agreements" etc, just walk away.

Yes, some companies have a threshold where managers can just "spend money". Some managers may even use that to support you. But taking any money changes the relationship you have with the user.

Right now, it's completely inside your control. Direction, Priorities, Scope, Pace, levels of effort etc. I'm a huge fan of getting paid, I write software for money, but make no mistake - taking money changes things.


Really: add a zero to your price. These companies burn millions on procurement bureaucracy. Make them pay for your misery.


Two zeros actually


fwiw the pricing page is buggy on mobile:

- the second 'start' button overlaps the list items

- tapping that button darkens the whole page but doesn't display anything new

- even before I get to that stage, it's not clear what you're selling at each price point


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

Search: