4 ms·

Somewhat related, I heard an interesting podcast recently by Corecursive about Sqlite and they mentioned how Android was so far ahead... excerpt: > This was ba

Somewhat related, I heard an interesting podcast recently by Corecursive about Sqlite and they mentioned how Android was so far ahead... excerpt: > This was back in 2005 or so, and we were in meetings with Android, and this was before Android was a thing... they had a prototype of their Android phone, and this was before iPhone... but we were debugging something with SQLite and we were plugging into the phone and we were running the debugger on a workstation which was pretty amazing. Nobody else could do that... here we were, we were debugging an application in GDB on a phone that was on the public network, and this was utterly mind blowing. Nobody at Motorola, nobody at Symbian, nobody at Nokia had anything close to that, and in that one moment, I knew that Android was going to be huge.

cpeterso 5y ago
Here's the podcast and a transcription: "The Untold Story of SQLite" https://corecursive.com/066-sqlite-with-richard-hipp/ https://corecursive.com/066-sqlite-with-richard-hipp/
AshamedCaptain 5y ago
Which is a ridiculous assertion, since I clearly remember running gdb targetting a Handspring/Palm way before the ARM transition, late 90s or the like. In fact m68k gdb was the only option available if you couldn't pay the big compilers. Damn gdbpanel. Not to mention that in 2005 Nokia already had the 770, which was basically a mobile desktop GNU/Linux device, with Gtk+, Gnome and everything. You could run gdb on the device itself.
- jsjohnst 5y ago
  > Which is a ridiculous assertion To the person who made the claim, it was the first, because of their limited view of the world. I always take claims of being first with a grain of salt for this reason. That said, I completely agree. I definitely used remote debuggers and on-device debuggers on mobile before Android/iOS existed.
  - cat199 5y ago
    > To the person who made the claim, it was the first, because of their limited view of the world. Then one would say 'it was the first i've seen' We have language for a reason. Despite what you hear on daytime talk shows about 'your truth', back in reality, actual truth is objective
    - ipaddr 5y ago
      Isn't that what everyone is saying. To say otherwise would mean you have a perfect understanding of something static. It is the first they know about. It may be defined differently. You cannot claim absolute truth on most things you can't can't claim absolute understanding of the language people use to express things. If you expect absolute truth in a variable world you're in for disappoinment.
      - neolog 5y ago
        I agree with parent. Speaker is a famous technical expert making public comments, people are gonna take him seriously. He should check that what he's saying is true.
- jeffbee 5y ago
  Yeah that's completely ridiculous. BlackBerry in the x86/C and ARM/Java eras could be attached to a workstation and debugged with Visual C or the later RIM JDK, setting breakpoints and single-stepping on real hardware live and on the air. Was debugging on the BlackBerry 857 over serial port attached to PC in 1999.
- wyldfire 5y ago
  > Which is a ridiculous assertion, since I clearly remember running gdb targetting a Handspring/Palm way before the ARM transition, late 90s or the like. From the description it sounds as if gdb was running off-target (same as it would for debugging Palm), except perhaps it connected via gdbremote-over-TCP instead of a 68k debug stub via RS232 or similar. And in this case, the phone's TCP stack is running over the phone's radio. > basically a mobile desktop GNU/Linux device From the sound of it, this device likely could have matched these claims.
  - jcun4128 5y ago
    The full couple of paragraphs is kind of lengthy to paste here but I think the main "coolness" is that while debugging, the phone rang. Not sure if in the other cases people are mentioning, the phone couldn't run. Also the development turn around time was much faster for Android's case apparently. > ...whereas the engineers that other companies, they had the big breadboard prototypes, the big full sized prototyping board, and the phone would run on that, and it was not connected to a radio so they couldn’t actually use it as a phone.
  - AshamedCaptain 5y ago
    > except perhaps it connected via gdbremote-over-TCP instead of a 68k debug stub via RS232 (he clearly says "plugged in" in the audio track, which does not hint to a over-the-air solution).
- lostmsu 5y ago
  Just gonna mention, that Windows Mobile had that feature too probably since its conception as Windows CE.
- gumby 5y ago
  Indeed, it was the handset Ericsson who paid for a bunch if that development in the 90s
- bostik 5y ago
  Disclosure: I worked, through a subcontractor, for Nokia from 2006 onwards. According to the people who deviced [no pun intended] the 770, it was never meant to be a product. It came originally as an internal prototype / developer machine, to make ongoing evaluation and development for maemo platform and its common applications easier. Nokia was obviously planning a polished consumer tablet. Someone high up decided the investment in hardware and supply chains for a developer-only system was not financially sensible, but if it was a product that could be sold to cover at least some parts of the setup cost, that would change the picture. (Incidentally, I believe, but can not prove, that it was also to gauge interest in tablet devices in general and to lay the ground for the future Nxxx series market.) Let's keep in mind, Nokia's consumer facing business was a PHONE company. We'll get back to that. The first consumer grade Nokia tablet was the N800. I vividly remember fixing its media player UI and working around various set-in-stone roadblocks in the GTK+ middleware. I also remember how our entire team were on a crunch schedule to get the feature completeness and bug rates to an acceptable level for going gold. The launch date had been set. We actually were told a few weeks before the internal deadline to stop working on all changes: our latest weekly progress report showed we had passed the launch requirement and the PMs didn't want to risk regressions from otherwise benign bug fixes. N800 had an FM radio. It had wireless internet and Bluetooth. It had a surprisingly functional front-facing video camera. It did not have phone capabilities. Neither did its upgrade, N810. The N8xx devices were incredible pieces for what they did, at the time. Techies loved them. The rest of the world didn't. Nokia was a phone company, and consumers associated the brand with the ability to make and receive calls. The first member of that family to have the missing capability was N900. Based on rumours and drunken vents in a sauna, the certification process to get that type of device with a baseband modem accepted was not a pleasant experience. ---- To this day, one of my proudest achievements in the maemo projects was to hunt down the root cause for a really nasty UI lag problem in the media player. Nokia was a design-heavy company, and someone in the design team had mandated a very specific type of GTK+ element to use in the problematic section of the player UI. The element did not support the necessary actions or properties, so in order to satisfy both the UI specification and the latency requirements, Nokia would have had to patch in a whole load of additional functionality to the element itself. This was a public API, and supporting that kind of bastardisation in a core library would have been an absolute nightmare. This bug was present in the run-up to N800. It was still there for N810, and it got bounced to me because I had argued about the infeasibility before the N800 launch. Once I had laid out the full technical details of the root cause, with a list of possible ways to fix it and the associated costs of each, I got the most satisfying resolution: "WONTFIX. We do not dare fix this bug."
kelnos 5y ago
I think that's just a case of someone being relatively ignorant of how things work in embedded contexts. gdbserver has been a thing since well before 2005, and remote debugging was not at all new, even then. If nobody at Motorola, Symbian, or Nokia could do (specifically) that, it just means that they didn't have GDB ported to their OS. But they absolutely certainly had other remote debugging tools; it's preposterous to suggest they didn't.
- pjmlp 5y ago
  I was at Nokia back on those years and also didn't get his point.

cpeterso 5y ago

Here's the podcast and a transcription: "The Untold Story of SQLite" https://corecursive.com/066-sqlite-with-richard-hipp/ https://corecursive.com/066-sqlite-with-richard-hipp/

AshamedCaptain 5y ago

Which is a ridiculous assertion, since I clearly remember running gdb targetting a Handspring/Palm way before the ARM transition, late 90s or the like. In fact m68k gdb was the only option available if you couldn't pay the big compilers. Damn gdbpanel. Not to mention that in 2005 Nokia already had the 770, which was basically a mobile desktop GNU/Linux device, with Gtk+, Gnome and everything. You could run gdb on the device itself.

jsjohnst 5y ago
> Which is a ridiculous assertion To the person who made the claim, it was the first, because of their limited view of the world. I always take claims of being first with a grain of salt for this reason. That said, I completely agree. I definitely used remote debuggers and on-device debuggers on mobile before Android/iOS existed.
- cat199 5y ago
  > To the person who made the claim, it was the first, because of their limited view of the world. Then one would say 'it was the first i've seen' We have language for a reason. Despite what you hear on daytime talk shows about 'your truth', back in reality, actual truth is objective
  - ipaddr 5y ago
    Isn't that what everyone is saying. To say otherwise would mean you have a perfect understanding of something static. It is the first they know about. It may be defined differently. You cannot claim absolute truth on most things you can't can't claim absolute understanding of the language people use to express things. If you expect absolute truth in a variable world you're in for disappoinment.
    - neolog 5y ago
      I agree with parent. Speaker is a famous technical expert making public comments, people are gonna take him seriously. He should check that what he's saying is true.
jeffbee 5y ago
Yeah that's completely ridiculous. BlackBerry in the x86/C and ARM/Java eras could be attached to a workstation and debugged with Visual C or the later RIM JDK, setting breakpoints and single-stepping on real hardware live and on the air. Was debugging on the BlackBerry 857 over serial port attached to PC in 1999.
wyldfire 5y ago
> Which is a ridiculous assertion, since I clearly remember running gdb targetting a Handspring/Palm way before the ARM transition, late 90s or the like. From the description it sounds as if gdb was running off-target (same as it would for debugging Palm), except perhaps it connected via gdbremote-over-TCP instead of a 68k debug stub via RS232 or similar. And in this case, the phone's TCP stack is running over the phone's radio. > basically a mobile desktop GNU/Linux device From the sound of it, this device likely could have matched these claims.
- jcun4128 5y ago
  The full couple of paragraphs is kind of lengthy to paste here but I think the main "coolness" is that while debugging, the phone rang. Not sure if in the other cases people are mentioning, the phone couldn't run. Also the development turn around time was much faster for Android's case apparently. > ...whereas the engineers that other companies, they had the big breadboard prototypes, the big full sized prototyping board, and the phone would run on that, and it was not connected to a radio so they couldn’t actually use it as a phone.
- AshamedCaptain 5y ago
  > except perhaps it connected via gdbremote-over-TCP instead of a 68k debug stub via RS232 (he clearly says "plugged in" in the audio track, which does not hint to a over-the-air solution).
lostmsu 5y ago
Just gonna mention, that Windows Mobile had that feature too probably since its conception as Windows CE.
gumby 5y ago
Indeed, it was the handset Ericsson who paid for a bunch if that development in the 90s
bostik 5y ago
Disclosure: I worked, through a subcontractor, for Nokia from 2006 onwards. According to the people who deviced [no pun intended] the 770, it was never meant to be a product. It came originally as an internal prototype / developer machine, to make ongoing evaluation and development for maemo platform and its common applications easier. Nokia was obviously planning a polished consumer tablet. Someone high up decided the investment in hardware and supply chains for a developer-only system was not financially sensible, but if it was a product that could be sold to cover at least some parts of the setup cost, that would change the picture. (Incidentally, I believe, but can not prove, that it was also to gauge interest in tablet devices in general and to lay the ground for the future Nxxx series market.) Let's keep in mind, Nokia's consumer facing business was a PHONE company. We'll get back to that. The first consumer grade Nokia tablet was the N800. I vividly remember fixing its media player UI and working around various set-in-stone roadblocks in the GTK+ middleware. I also remember how our entire team were on a crunch schedule to get the feature completeness and bug rates to an acceptable level for going gold. The launch date had been set. We actually were told a few weeks before the internal deadline to stop working on all changes: our latest weekly progress report showed we had passed the launch requirement and the PMs didn't want to risk regressions from otherwise benign bug fixes. N800 had an FM radio. It had wireless internet and Bluetooth. It had a surprisingly functional front-facing video camera. It did not have phone capabilities. Neither did its upgrade, N810. The N8xx devices were incredible pieces for what they did, at the time. Techies loved them. The rest of the world didn't. Nokia was a phone company, and consumers associated the brand with the ability to make and receive calls. The first member of that family to have the missing capability was N900. Based on rumours and drunken vents in a sauna, the certification process to get that type of device with a baseband modem accepted was not a pleasant experience. ---- To this day, one of my proudest achievements in the maemo projects was to hunt down the root cause for a really nasty UI lag problem in the media player. Nokia was a design-heavy company, and someone in the design team had mandated a very specific type of GTK+ element to use in the problematic section of the player UI. The element did not support the necessary actions or properties, so in order to satisfy both the UI specification and the latency requirements, Nokia would have had to patch in a whole load of additional functionality to the element itself. This was a public API, and supporting that kind of bastardisation in a core library would have been an absolute nightmare. This bug was present in the run-up to N800. It was still there for N810, and it got bounced to me because I had argued about the infeasibility before the N800 launch. Once I had laid out the full technical details of the root cause, with a list of possible ways to fix it and the associated costs of each, I got the most satisfying resolution: "WONTFIX. We do not dare fix this bug."

kelnos 5y ago

I think that's just a case of someone being relatively ignorant of how things work in embedded contexts. gdbserver has been a thing since well before 2005, and remote debugging was not at all new, even then. If nobody at Motorola, Symbian, or Nokia could do (specifically) that, it just means that they didn't have GDB ported to their OS. But they absolutely certainly had other remote debugging tools; it's preposterous to suggest they didn't.

pjmlp 5y ago
I was at Nokia back on those years and also didn't get his point.