The new epoc(h)

Way back in March at CeBIT, Symbian announced that they had shipped the latest version of epoc to their licensees; at the same time, Psion Industrial, now Psion Enterprise (the name change is because they now support Java) announced two machines, known under the project code names of "Jedi" and "Quantum". Ericsson also announced at about the same time what appeared to be a rebadged Series 5. It turns out that all of these machines, as well as the just launched Series 5mx, run the announced Symbian product - ER5 (epoc Release 5).

Earlier versions of epoc, ER1 and ER2 were shipped on the Series 5 and GeoFox; ER3 and ER4 were never released in full form, but included the incremental releases of the Messaging Suite (EPOC Message) and PsiWin (EPOC Connect) that we have seen since then. ER5 includes significant changes to these components, as well as new applications, and a new Java SDK. The most immediately obvious of which is that it supports colour, but only (so far) on the Psion Enterprise "Jedi".

The messaging suite components, Email and Web, have been revised extensively. Mail includes SMS support and folders, for example, and Web now supports frames. Mail uses an add-on arhitecture to support IMAP4, which should also make third party customisations feasible. SSL is still not supported, and PGP is supposedly under development by a third party.

A significant new application is Contacts. This replaces Data as the principal address book, and is supported by EPOC Connect/PsiWin and Email. Basic fields are phone, name, address and email, but custom fields can be added per record. Contacts also support CLID (caller id lookup). The standard applications have had minor improvements: Sheet now sorts, Agenda has a busy view, Word supports both templates and outlines, and System supports multi-file beaming and wall paper. The shell can also now directly launch files that are in non-native formats in the appropriate application, such as Microsoft Word files into Word. Jotter has reappeared, and found its way onto the button bar.

PsiWin 2.3 consists of many bug fixes to PsiWin 2.2, speed improvements, connectivity for the new Contacts app, and for Email. Email will synchronise with Outlook and cc:Mail; Contacts syncs with the usual suspects, and uses a new architecture that uses vCard 2.1 as a base. Agenda sync also uses these formats: vCalendar 1.0; others of these formats should also be supported - vMessage (see http://www.imc.org). This should mean that business cards can be beamed between a Series 5mx and a Palm III, for instance. Applications that access files have been changed to use a client/server architecture, which means that synchronisation no longer needs to close applications to operate.

Java support is for 1.1.4, including AWT, and is certified "100% pure Java". This is a complete implementation that should allow properly written cross platform Java applications to run. Note that 1.1.8 is the current release of Java 1.1, and Java 1.2/2.0 is available now, and is required to support Swing, which I would expect most modern Java applications to require. Speed of execution and of loading has been optimised; Java support required the ROM to increase in size from 6 Mb to 8 Mb, and adds a 1 Mb overhead to RAM when in use.

Another under the covers change is support for ETEL, a telephony server API, with specific support for GSM.

Future versions of epoc will support more RAM processors, and the Motorola M.CORE CPU. Other possibilities being thrown around include: unicode, WAP, Bluetooth, scaleable fonts, encryption, remote synchronisation, device to device synchronisation, Starfishs's TrueSync synchronisation architecture and possible USB support.

Apparently Psion plc are considering releasing this version as a ROM upgrade for the Series 5, although release isn't certain.

Symbian will support a number of different device formats for epoc; these are called "Device Family Reference Designs", or DFRDs for short. All DFRDs are for either smartphones, which are keypad based with a relatively small display, or communicators, which have larger displays and keyboards. The four DFRDs announced so far are:

Emerald - a pen based smartphone with a 320 x 120 display;
Sapphire - a keypad smartphone with 200 x 200 display;
Quartz - a pen based communicator with 240 x 320 display;
Crystal - a keyboard communicator with 640 x 200 (half VGA).

The Series 5mx and Ericsson MC218 are based on the Crystal DFRD.

The Series 5mx will look very much like the Series 5, but with ER5 and 16 Mb RAM, and the ROM increased to 8 Mb. The CPU is a 200 MHz ARM. It will be in the shops before the end of June; a price isn't know at this time (but has probably already appeared in Horizons and the Reviews section).

The Psion Enterprise "Jedi" is identical to the Series 5mx internally, but with a colour VGA size display. The "Quantum" is a ruggedised, keyboardless model with a half-VGA display, but is otherwise identical. The Ericsson MC218 should ship in July or August, and has some additional or changed applications; the changes are most as for the Series 5mx, as a consequence of using ER5, but will also include a WAP application.

At the moment, it looks as if Symbian have all the smart money on taking the mobile phone market. The alternatives are Palm and Microsoft: Palm have already produced a mobile phone with Qualcomm, but their PalmOS doesn't show the voice oriented features and ease of development of Symbian's far more mature operating system. Microsoft seem to have missed the boat by ignoring the mobile phone manufacturers, and CE has lacked the critical real time features since its introduction. However, Microsoft have also produced a mobile phone, jointly with Siemens, have been promising real time for the last several years, and are rumoured to have suddenly taken an interest in this market, showing prototypes of new CE models to phone manufacturers.

However, Symbian have not yet delivered a phone based on their OS; the Ericsson R380 isn't due until next year. The phone manufacturers are notoriously uncommitted to any single operating system, and will readily license a new one for a single product release. This means both that it is more likely that they form an alliance for expediency with Microsoft, perhaps for a single model, and that they will not be swayed by any bogus arguments that the Windows user interface is desirable on small devices: it isn't. In this market, software developers really shouldn't count their chickens until they are hatched, and the eggs are individual mobile phones in the hands of consumers.

The main reason that Psion have had a shot at this market at all is that the usual American practice of jumping the gun on standardisation has lead to a fragmented US market, but a global market unified mainly around GSM. Europe has had close to 100% coverage for several years, and even now an American network is proud if it can boast better than 80% coverage - which early Orange and One-to-One users in the UK know is almost as bad as no coverage at all. As a result, the European market has been able to support local developers for some time, and the American software giants have been blind to the opportunity.

The current non-wired PDA market is small in mobile phone terms: 2 to 5 million as against several hundred million units. But in the immediate future most phones will take on PDA attributes, and most PDAs will develop wireless networking using largely mobile phone technology. New wireless protocols for both local area (Bluetooth) and wide area (several standards are being implemented at the moment) have the megabit capacities that will make this a convenient reality, instead of the awkward, slow, multibox shuffle that is required today. These PDAs will be sold as phones, and will be made by mobile phone manufacturers. So IDC reports that postulate a mere 20 million market for PDAs in five years time may be both right and wrong, and certainly demonstrate the futility of market surveys in an evolving market: PDA-only units may sell that poorly, but mobile phones will do all that PDAs currently do (and, significantly, will run the same OSes as current PDAs). According to Ericsson, the phone market is currenly in excess of about 200 million units, and is projected to reach 600 million in 2001.

PalmOS is great for current PDAs; it has the virtues of simplicity and compactness. However, it hasn't yet been adapted to support proper real time operation, colour, or voice telephony. Palm understand, better than most computer manufacturers, the mobile phone market, and often compare their marketing to that of a phone manufacturer; but they still have to deliver a technology appropriate for the application. Microsoft don't seem to get any of this at all, but their usual 'Chinese water torture' approach to incremental development, and sheer persistence, could easily get them there. Symbian, however, have a released operating system that has both the stability of long development and testing, and an understanding of the requirements of its co-owners. Their approach to the support of multiple form factors is far better, and they already have toolkits for the main technologies: telephony abstraction, GSM, WAP and so on. This is the golden opportunity for them.

Words and design by:
Paul Lynch
Last updated: August 19, 1999