Was that quote from NeXT aimed at sysadmins, so that the whole system was to be managed via GUI?

The question is what people require in a Maemo tablet. An input method provides just that, an input method, but it always comes with a price: physical weight, size, power usage, software has to support and be optimized for it, and so on.

Something like a TrackPoint also adds value, and for me it works very well compared to a mouse. Better than a dpad. But for some purposes it'd be worse than a dpad.

So we established each input method has their place and their positive and negative aspects we must also not forget a new input method or new application or new device takes time to get used to. The fact it is not what a user is used to does not mean it is bad. So it might for example take time till one is used to a virtual keyboard, after which the performance of it is more acceptable than it was in first experience. The hardware used (e.g. screen size, type of screen) and the software implementation of the input method (e.g. virtual keyboard) heavily influence the end result as well.

I think what a corporation lika Nokia has to do (and does) is what they did in with their phones for ages: they need to draw profiles of users, profiles of user cases, and optimized combinations of hardware. [EDIT]This is quite difficult! Tough decisions are made![/EDIT]

Hence the question of what people require in a Maemo tablet isn't very descriptive. What would be better feedback is a compilation of in depth responses to a poll like this, and have those people who'd like a 'N800 succesor' (a traditional yet fast touchscreen device with not many input methods?) team up in this regard (in Maemo Brainstorm).

Tablets are also used to remote control other devices via console (SSH, ...) or GUI (RDP, VNC, NX, BlueMaemo, ...). But afaik there is no client for Maemo which has this zooming in and out feature LogMeIn Ignition for iPhone/iPod touch has. Which also has a slick interface, optimized for iPhone/iPod touch. They have more advanced products too which come with a heavier pricetag. Yet, this is not the primary function of a tablet.

That is what people here tend to forget. You cannot simply easily do 1:1 port of existing *NIX software to an embedded *NIX platform and be done with it; the UI really needs work, especially if you cater to the masses. Actually not only the UI needs work there are other issues too like bloat related issues or all kind of dependancies which are a given on a desktop but are not preferred on an embedded system. A great example is the porting of a Gecko browser to ARM. So, the Maemo software must all be optimized for the device. You simply cannot have your UI optimized for both stylus and finger. You cannot have your UI optimized for keyboard and virtual keyboard. Or at the very least, this is very difficult.

This is inherent to embedded products. You see that in Nokia's products too. Every device runs their own SymbianOS which besides a standard base is optimized for that purpose. Compare with Maemo versions. Now, it is inherent to any product, even laptops run different software than PCs. Yet some people want to run all their PC software on their tablet. People run BitTorrent on their embedded devices, Sure, you can do that, but all that hashing costs precious CPU cycles, RAM, and I/O to/from storage. And it costs bandwidth. You're free to do this, but you are not the target market Nokia will aim for because there are not enough people who use a device like you do. Because other people simply run their BitTorrent client at their home desktop and use a client on their embedded/mobile device to control that client. Much simples, costs less resources, and the desktop has its own internet connection and resources. You say is too bloated to keep BitTorrent client 24/7, fair enough, people run this on embedded ARM devices as well. Like OpenWRT. But these don't have a screen. If they have something like UPnP-AV you could afterwards simply access it with your Maemo 5 device. Anyway, such is a niche, just like remote access is a niche. Unless a very slick application like that is developed for Maemo. So instead a corporation which develops Maemo for their end product has to define certain user cases and purposes. Their own base + functionalities do that, their advanced and strong framework allows users to extend it, and they keep in mind something like the Fremantle Stars. What you cannot expect is that all kind of input methods and wicked purposes are kept in mind. Because...

An embedded system is a special-purpose computer system designed to perform one or a few dedicated functions, often with real-time computing constraints. While a smartphone, DAP, tablet is not strictly an embedded device anymore (it does more and more) it will not and cannot ever fullfill the very same a PC does. If you aim for that, your aim might be noble but far too high, and in a way disrespect the advantages of the device. You want a steady evolution towards such path. Not a revolution.

You're more describing all the features you'd like on your Maemo tablet. That is a different topic. Try to list the few features which are most important to you. E.g. dealbreakers. BusyBox you can manually replace, but comes with a cost.

missing GPS as a poll option.

Some things I assumed would be in there, since the limit on the forum is 10 items in the survey. Those were things that aren't very controversial: USB Host+OTG, GPS, Stylus.