[Gta04-owner] gemini PDA

H. Nikolaus Schaller hns at goldelico.com
Wed Mar 8 16:58:39 CET 2017

Hi Thomas,

> Am 08.03.2017 um 15:43 schrieb thomasg <thomas at gstaedtner.net>:
> On Mon, Mar 6, 2017 at 3:31 PM, Belisko Marek <marek.belisko at gmail.com> wrote:
> Seems that Indiegogo campaign was successful and this device will be
> shipped to users. Looks like competitor for Pyra:
> https://www.indiegogo.com/projects/gemini-pda-android-linux-keyboard-mobile-device-phone#/
> I think there are a few points to mention:
> 1) Even if a crowdfunding campaign is successful, it doesn't mean that THE product (or even A product) will be shipped to donors
> 2) The company behind the campaign does not seem to have experience with hardware of that complexity level at all
> 3) they claim to have a fully functional prototype device, but it is not shown. in fact nothing is shown but rendered models
> 4) the "reward" tiers are surprisingly low for a device of that complexity, produced with quite limited budget. USD 500,000 is not enough for a mass production run of this device, and the average reward tiers are too low to pay for a device with a production run under 10k units.
> 5) the platform is based around a Mediatek SoC, a company with a quite questionable track record in GPL compliance
> 6) they do claim linux support, which seems unlikely since they are dependent on Mediatek source code which is android-only (i. e. the built in LTE modem cannot ship linux

I agree that Mediatek SoC's are not he first choice for GPL code...

>  drivers, but will depend on a closed source android RIL blob, the Mali GPU will only be usable with a proprietary android "SurfaceFlinger" display driver, no X and no wayland support; amonst other problems)
> 7) there is no evidence of any open source ambition or experience (which matters to this community I think)
> 8) there is very little transparency. No information is given how they expect to actually produce the devices. No information to software or the prototype is available.
> Let's estimate what the components will actually cost:
> Mediatek sells its Helios X20 dev board for USD 199, this is a bare bones SBC without any extra features, and likely close to production cost (as this very likely is a sub 10k production run as well). The SoC alone is reported to be around USD 30, if you buy 100k or more.
> The advertised display is LGs newest and by far most expensive (small) display to date which will easily be above USD 50 for 100k units.
> On that basis, I'd estimate large quantity orders for components alone will reach USD 150, for medium orders it might be almost twice that.

Yes, that fits our own experiences with the GTA04.

> This leaves a very tiny margin for production problems, if any.

Let's hope they don't fall into a pit like we did.

> To make it clear: I'm not saying that they won't succeed or can't succeed. In my opinion it's just unlikely.
> There's always good reason to be skeptical with crowdfunding, even more so with highly ambitious projects, and much more so with relatively small budgets.
> Many crowdfunding campaigns fail. _Most_ such ambitious campaigns fail.

In my observation those campaigns succeed where development is already almost done and the organization behind the idea could do it on their own.

Then, crowdfunding platforms are additional marketing tools to easily attract customers, combined with a sales channel because it is even paid.
Normally, marketing and sales go separately (e.g. Google Ads and some online shop or Amazon).

> In my amateur opinion, producing a device of this kind on a 500k budget for USD 400 per device cannot be done.

Thanks for sharing your interestingly detailled view,

More information about the Gta04-owner mailing list