As i checked on google, pagespeed measures mobile and desktop sites using "Emulated 3G network". As per some GSMA report from 2016 that 75% or users are on 2G/3G networks.
Thats almost 2019 year and many users are on 4G now, so reports are useless.
Can someone tell Google that either they should use "No throttling" or add switcher for 4G network.
Per GSM report for 2017, 71% of Global users are on 2G/3G network. That number drops to 32% for North America and 59% for Europe. While we are definitely moving toward 4G dominance (predicted by 2025), it seems too early to ignore such a big part of consumers.
If your business have customers buying from 3g networks then take page insights data into consideration. Otherwise just don't look at them, they are misleading for most use cases. Google want us to believe we live in Africa all together. Note they are also emulate a moto 4g, a mobile phone from 2016 for their tests. Another fail for the modern era.
Related
Looking for some idea on security related to medical devices.. like there can be OEM(s) who manufacture the devices or can be hospitals buying it out from different vendors for setting up their infrastructure. How do we really identify whether a device is IoT enabled and security assesment needs to be done. Like example, if we look at Johnson & Johnson or Baxter .. they build medical devices but how do we know it's an IoT and related security to be addressed. Any kind of information and guidance would be of great help. Thanks :)
I would take a look at alibaba as many companies source china for manufacturing. Typically knock offs will occur(or its just plain OEM'ed from china) and will prob show up on places like alibaba. Within the listings of alibaba you can see if they list or you can email the supplier and ask if they use dhawa, hisilicon or any of the other IoT boards in their product.
I have a number of required business cases for HoloLens that require the device to understand a general geolocation, such as the current wearer longitude and latitude within 10 meters or so, as well as sending location information to and from an endpoint during various processes. Users WILL have a mobile device with geolocation capabilities that could assist in the process if necessary, and could also be used as a WiFi hotspot.
Is this a reasonable and reliable use case for HoloLens? Can apps be created that use geolocation and maintain connectivity during an experience, either on their own or with real-time communication to and from a mobile device that has these capabilities?
Yes - definitely is a realistic scenario. I've done some integration between a blue-tooth GPS and the Hololens. Let me know the particular device you're looking at and I'll see if I can get it working with the Hololens. //Lance Larsen (Microsoft MVP) - www.lancelarsen.com
I have a couple of questions concerning BLE beacons:
1) Are beacons based on nRF51822 chip the best solution? Or are there any other chips better than nRF51822? I want to take up BLE beacon development and struggling to find the right hardware for these needs. As a novice developer I want the beacon to be as cheap as possible in order not to waste money in case of a failure.
2) Is it possible to buy pure Eddystone beacon (not iBeacon)? The reason for choosing Eddystone is that Eddystone is capable of broadcasting URLs that are essential for me.
The second question stems from my failed attempts to find a pure Eddystone beacon on Chinese electronics sites like alibaba.com or aliexpress.com where the only firmware available is iBeacon. But iBeacon is not an option because it can't broadcast URL the way Eddystone does.
Apart from the above questions It would be great if someone wrote a quick guide for taking up BLE development with Eddystone and covered basic topics like: chip to use, beacon model, best website to buy beacons at, etc.
Thanks in advance,
Pavel
1) I've worked with Estimote beacons and Chinese beacons from Amazon and in my opinion, they do not differ in terms of accuracy too much. Especially for prototyping, I'd buy cheaper ones to test if your use case can be satisfied with BLE beacons. If it is too inaccurate with Chinese beacons, chances are that it won't work with more expensive ones either.
2) Why do you need the URL broadcast? If the app is going to use the url, it would have to be connected to the internet. Therefore, you can just query the beacon's IDs to a web service to get back an URL and use that. Personally, I think this is a better approach as you can configure the web service from anywhere to change the url for beacons where as if you want to change the URL of the Eddystone, you have to go to the beacon to configure it.
The nRF51822 is a common implementation, is flexible, well understood and can be very inexpensive. Be aware though that development costs, add on circuitry for power and/or peripherals, and packaging can easily eclipse the Bluetooth chip when you get to production cost savings.
If you want to buy an off the shelf beacon, most models supporting Eddystone also support iBeacon, simply because supporting both adds no additional hardware cost. Newer Radius Networks and Estimote beacons all support both. And, yes, cheaper generic Chinese suppliers often have bulk manufactured inventory from before Eddystone existed at only support iBeacon.
I'm looking for a development board for WinCE, preferably with the following connectors:
VGA/DVI for screen
USB for kb/mouse
Ethernet (10Mbit is OK, 100 would be even better ;-) )
Storage (eg: SD/MMC/CF cards)
I'm looking for this, because I want to attend the Imagine Cup , Embedded Systems design. And I'm looking for a challenge, ofcourse ;-).
I found a board called HawkBoard (http://www.hawkboard.org), but I want to know if there are alternatives for about the same price (max $200 incl. shipping/customs/taxes). I need the BSP, ofcourse ;-).
Thanks!
Yvan
The eBox 2300/3300 is a low cost device that has been used for past Imagine Cup challenges. It's x86, comes with a BSP, is in an enclosure, is low cost ($195 or $255) and meets most of your needs.
Of course CE has forever shipped with a reference BSP for x86, so you can usually get just about any x86 device working, though most other devices tend to be pretty big physically.
There are also open source CE BSPs available for the BeagleBoard and GumStix.
The BeagleBoard is a good ARM system that costs $150. It has all the features you want, from the product page:
DVI-D for connecting digital computer monitors
Compatibility with a huge collection of USB peripherals including hubs, keyboards, mice, WiFi, Bluetooth, web cameras, and much more
MMC+/SD/SDIO interface for memory or wireless connectivity
S-Video out for connecting your NTSC or PAL television or wearable visor
Stereo audio in and out for a microphone and headphones or speakers
Power via typical USB chargers for cell phones from your laptop, from an automobile adapter, from batteries, or even from a solar backpack
I never tried Windows CE myself, but a google search suggests it's possible, and there's even a video on YouTube.
Another option is Gumstix, and there's a project on codeplex to help you get Windows CE on it.
I guess it also depends if you want to start something for fun or rather serious.
Check some of the big manufacturers homepages too, so you can see what specifications could
help the realisation of your ideas or set limits to them.
What helps for shure is if you choose a platform, that can be configured easily and its much more fun if you start with something that can work out of the box too.
Here an example, like you can find it on a lot of manufactureres homepages on the web.
http://www.toradex.com/En/Products/Colibri
Colibri modules feature state of the art technology, and aimed at low power systems that require high CPU performance. They are equipped withall the interfaces needed in a modern embedded device.
Colibri T20 features include:
· Dual Core ARM Cortex A9 Processor
· 2 x 1.0GHz Processor Speed
· Full HD 1080p Video (Encode / Decode)
· Powerful 3D Rendering Performance
· True Dual Display
· Up to 12 Megapixel Camera Sensor Support
And if this is to much there are Colibri modules in the family which are in a lower price range with adapted specifications that might fit you needs better.
What I personally think is important, that you know that you can get good support and talk to the right people in person as well.
Give it a shot and good luck.
Regards
Any old PC that has a video adapter that is VESA compliant and a network adapter based on NE2000, realtek RTL8139 or 3COm 3Cx5 can be used as Windows CE target device using the CEPC BSP.
If you have an old 486 or more powerful PC lying unused somewhere in your garage, it's time to resurrect it! :)
I have very easy question, but I simply have any idea of the answer.
I have developed a small mobile-application using java, for my nokia.
The problem is that when installed on my samsung the application simply crashed.
Then I tried on my other nokia but different model, and I didn't got the normal behavior.
So my question is, does anyone have any idea how companies that develop mobiles applications/games test their software.
Does they have to have all models for all mobiles phones??
Companies that target many phones in many countries usually only let you install the application on your phone if they recognise your handset User Agent in the HTTP headers of the request to download the .jad or .jar file.
There are multiple ways to test an application on many handsets for many mobile network operators.
From simply buying the phones, to establishing commercial parternships with handset manufacturers and mobile network operators, to having a Device Anywhere account.
I don't know if you need all models of all phones. But you will definitely need separate test (and probably different builds) for different phones regarding:
MIDP version
Screen Size
Input Devices
Speed & Memory
Java, in this case is, WOTA (Write Once Test Anywhere) instead of WORA (Write Once Run Anywhere). :-)
Phone specs and Java implementations vary a lot, but within each manufacturers range there will be groups of phones that share the same specs and implementation.
I used to work at a company making J2ME games, what we did there was test on every handset we released the game on, but we had 2 types of test - Complete and Compatability.
We would adapt a version of the game for a specific phone, eg Sony Erricson K800i, and have it thoroughly tested according to the Complete Test spec.
Once that had passed, we then used that build on phone known to have similar specs and good previous compatability with other games (we kept a database of specs and compatability records), eg Sony Erricson W910i, and submit it for a compatability test, which was a bit less thorough and a bit quicker.
Once you've been doing it a while you get to know the capabilities of phones and which phones you could use the same build on, but there is often a bit of guesswork involved :) Sometimes you get matches you wouldn't expect, and sometimes a match you would expect to work doesn't.
Edit: I was going to post this as a comment, but I can't (because i'm an SO noob :), out of interest, what phones are your Nokia's and Samsung?
I can't remember many specific handset names, but here is a quick rundown of compatability across manufacturers:
Sony Erricsons are generally excellent - if it works on one, it will likely work on all SE handsets with the same resolution.
Nokia's are generally good within a certain smaller group eg N95 builds work well on most nokias with the same res that were released after the N95, but some handsets are a bit of a pain.
Samsungs are pretty bad - the J2ME implementation on most is flawed (Hide/Show Notify methods not being called is an example), and the memory and speed are typicly a bit crap.
Motorola phones are not great, but are generally quite compatable with oneanother. Same goes for LG, although their more recent models are much better.
Testing is one of the most labour intensive part of mobile phone development. Typically a company might simply buy a lot of different phones to test on for real, or target a particular subset such only as Series 40 Nokia phones.
But alternatives exist out there where you can remotely deploy your app to phones, such as Nokia's Remote Device Access Services.
One way that might limit the problems is to target J2ME MSA (Mobile Service Architecture) compliant phones, where MSA attempts to reduce variations in vendor implementations of J2ME.