1. 46
puri.sm

The crowdfunder reached its goal after 6 weeks, and still has 2 weeks to go. The stretch goals are as follows:

• 4 million \$: VOIP phone number, call-in, call-out features
• 6 million \$: Reverse engineering faster WiFi/BT firmware
• 8 million \$: Free encrypted VPN tunnel service for all backers for 1 year
• 10 million \$: Create an isolation layer to run Android applications on Librem 5
1.

2. 5

I’ve been noting down the crowdfunder’s progress every so often – here, have a plot of the crowdfunder’s progress. You want the second image in the gallery.

• Sometimes I’d measure again quite soon after the previous measurement, forming clusters of two or three or even five measurements. These observations are obviously correlated, and I didn’t want the linear model I used to give special weight to points that are close together.
• In general, I used a linear model because I couldn’t find a time series model that handles irregularly-spaced measurements. Solving this problem should also solve the previous problem.

In the end, because this is an extremely unimportant activity, I used a weighted least squares model that weighted each point by the sum of the distance two its next and previous neighbour. That way points in the middle or on the edge of cluster get de-emphasised, and the sum of the weight over the entire line does not depend on the number of points. Anybody got any suggestions on doing this properly?

1. 3

The CSV data, should anybody want to mess therewith. Free of injection attacks, I promise.

date,amount
2017-10-01 22:36:41,901446
2017-10-01 22:37:04,901446
2017-10-01 22:51:57,902045
2017-10-02 18:41:43,921294
2017-10-02 21:03:27,924888
2017-10-02 21:04:32,924888
2017-10-02 22:16:50,924908
2017-10-03 11:48:46,942419
2017-10-03 13:32:17,948128
2017-10-04 10:35:51,967434
2017-10-04 11:51:45,969271
2017-10-05 11:50:08,1022619
2017-10-05 13:11:02,1023517
2017-10-05 13:38:14,1024775
2017-10-05 15:09:45,1030546
2017-10-05 15:11:09,1030546
2017-10-05 17:50:28,1042825
2017-10-05 17:51:27,1042825
2017-10-05 17:52:06,1042825
2017-10-06 15:02:58,1135245
2017-10-06 15:05:33,1135245
2017-10-07 22:38:52,1314020
2017-10-08 11:39:44,1344817
2017-10-08 14:45:35,1353403
2017-10-09 11:20:27,1402860
2017-10-09 15:11:07,1407414
2017-10-10 10:02:12,1543038
2017-10-10 19:40:47,1563404

2. 3

Please, please can we refrain from consumer product news here? Even for projects we like?

1. 25

This isn’t yet another iPhone review - it is much more in line, interesting, and relevant, to the readers of this link aggregator, than any bog-standard consumer product.

I’m glad to have found out about it here as, since I’m not following the project directly, would have probably missed the news altogether.

1. 4

This is how you get overrun by yet another iPhone review.

Everybody has some product that is they think is “more relevant” to users of this site, and if we post those, we set precedent for that, and then we get overrun. This happens on HN all the time.

If you want product news, go look at a product news site. News is the mindkiller. Ads are the mindkiller. Content marketing is the mindkiller.

If this was a long-form article on how they structured their kickstarter, or on how they did market research, or on how they adapted their phone, that’d be one thing (and one better suited to barnacl.es at that!)–but it’s not. It’s a straight 3 paragraph press release.

(I’m willing to bet you aren’t even going to buy one of these phones.)

1. [Comment removed by author]

1. 2

There does seem to me to be some qualitative difference between the release announcement I posted with links to direct engineering resources and project history, and this press release.

Care to elaborate on your observation?

2. 6

It’s pretty rare that there is a libre product in that space, so I don’t see the jump between allowing this and being overrun by iPhone news (unless you do mentally categorise them in the same group).

If you want product news, go look at a product news site. News is the mindkiller. Ads are the mindkiller. Content marketing is the mindkiller.

I understand, you want the news that fit your definition of news. But we’re a group here. That’s why you found out you can effectively derail any news you don’t like by starting a meta-discussion about not wanting that content here. I’ve never seen your doom-saying turning out to be true, though. You are taking the same approach for years, and I have yet to see this page being overrun by corporate marketing or ads.

Fear is the mind-killer, nothing else, by the way ;).

1. 4

Lobsters isn’t a news aggregator–it’s rather a bit deeper and more durable than that.

You are taking the same approach for years, and I have yet to see this page being overrun by corporate marketing or ads.

There may or may not be a correlation there…

2. 8

Just filter the hardware tag - you’ll be much happier :^)

(I’m willing to bet you aren’t even going to buy one of these phones.)

I’m not even going to dignify that with a response.

1. 5

While I agree with @angersock that we should avoid product announcements on this site, I think this product is an edge case - it is unique in the sense that it is focused on bringing FOSS to a place that has been elusive for a long time (mobile phones).

Your suggestion of filtering the hardware tag is ridiculous though - there is a lot of interesting hardware hacking on there that is not just a link to some product launch article.

3. 1

1. 14

To be fair, that’s how you end up as reddit.

1. 4

Reddit and lobste.rs are very different products, you are basically comparing apples and oranges.

Even when taking this wide stretch, lobste.rs would be more akin to a subreddit.

1. 3

An online forum with tree-based comments, with upvotes and downvotes to sort content. They’re pretty similar. Every growing online community is subject to the Eternal September problem, and I don’t think plain voting is enough to solve the issue.

Instead of actually solving Eternal September, Reddit sidestepped the issue by making it very easy to create new subreddits. Those small communities can be rather high-quality, whereas the popular front-page subreddits are full of “empty calories” - flashy content that’s easy to consume and appeals to a lower common denominator.

I enjoy angersock’s feedback, because he’s consistent about calling out “fluffy” content. Relying on votes isn’t a good way to address fluff, because fluff is easy to upvote. It needs to be called out.

2. 1

While I would normally agree with you, Lobste.rs is invite only and isn’t open to any random troll to join. This is why I love lobsters because the signal to noise ratio is much higher. People care and the votes reflect the culture based on this selection process.

I wonder if there is a further business model here, similar to metafilter. Keep invite only, but only allow votes for people who actually pay a small subscription fee to even further disincentivize reddit like behavior.

2. 1

Is it realistic to build a phone and software with that money? Many have failed with more resources

1. 4

They have built laptops with much less crowdfunded, so there’s a reasonable chance that they will ship the Librem 5 too. I imagine crowd funding is not the only source of money in their budget (they did spend a year and a half on research prior to launching the campaign, after all).

1. 2

Yeah, something definitely isn’t adding up. The \$4M and \$8M stretch goals seem like snakeoil too.

1. 2

I should say it should be possible, for the reasons Algernon put forth. For reference, these were the targets of previous Fairphone crowdfunders, which both succesfully reached production and had follow-up batches made. They also both included a degree of OS development (Android customization) – Fairphone OS’s home screen / launcher is still the nicest I have used to date, and not just because it’s Google-free.

• 2012 (Fairphone 1): 1.5 million € (5000 pre-orders at 325 € each). Selling 5000 phones was enough for them to start producing their first batch of 20 000 phones.
• 2015 (Fairphone 2): 7.9 million € (15000 pre-orders at 529 € each). Not sure what the production batch size was. Cost breakdown.

(Edits: formatting)

1. 0

It depends. Will they have a brick with a capacitive touch screen and some modems in it, that runs on battery power? Almost certainly yes. Will it be something that anybody wants (for some reasonable definition of “anybody”)? Almost certainly not.