Comments on: PalmSource: One on One with Mike Mace

Thanks,

\\ AC

While reading the interview, I kept thinking isn't the title of CHIEF COMPETITIVE OFFICER ridiculous ?!? Talk about too many CHIEFs and not enough indians. Palm was and is still bitten by position inflation. At a dot.bomb where I worked long ago, they had such inflated titled such as CHIEF PEOPLE OFFICER (i.e., human resource dork) and CHIEF CREATIVE OFFICER (the most uncreative person I've met in my life).

Everyone wanted to be a CHIEF and not enough indians to do the work. Net effect: the company tanked. Stock dropped to the nether realms. And the company CEO is still carrying on his affair with his executive assistant (that was not a well kept secret either).

I think I've found my new tag-line!

I'm serious.

Your feedback is much appreciated.

If I could just have this, I would be happy as a clam-

Is there any serious problem in the underlying OS or 3rd-party app compatability that prevents this?

1. While I understand logically the need to split the OS releases up into more bite sized bits, the sad truth is that Palm needs some sorely need obvious to the end user improvements right now, not later. Sure I understand all the reasons why that might not be possible but that doesn't change the fact it is needed. By the time OS 6 or whatever it will be called arrives with a more updated GUI, multithreading, memory protection, native ARM apps not through PACE, and updated PIM apps it will be mid 2003 and where will Palm's competitors be then? I don't know where PocketPC will go, and it maybe it will still continue in more expensive, larger sized handheld route in which case Palm is still same... but then again maybe we will see an improved GUI, iPAQ the size of an m505, all for $399 by mid 2003 too.

2. It is short sighted some of his statements on the PIM. We do not want ActionNames or Datebk4 functionality in the core OS, but improved and updated PIM functions need to be one of the highest priorities for any of the new OS's. The core apps of the Palm haven't been updated since 1997 in OS 2.0 of the PalmOS and their inherent database structure is actually holding back even the aftermarket apps like ActionNames etc. because of it.

3. I am skeptical about the 6-9 month claim between OS releases. All I can say is I hope Palm and the other companies making PalmOS units reward those of us who buy bleeding edge OS 5 units with quick and timely flash updates for no more than an nominal (read under $40-50) fee when these updates come out. My biggest fear in buying an OS 5 unit is not finding that in 6 months a much faster unit comes out, but rather than a major OS release with the things I finally want arrives but I can't have it without buying a whole new unit.

Im working on a mass transit scheduling app and my first thought and hope was to use Java on PalmOS. I found IBM's VAME to be the best implementation but I couldnt afford the $12/copy runtime licence fees. I've since moved to straight "c" development with the gcc tools. I've been picking up more and more of the PalmOS API.

My conclusion so far is that the PalmOS 3.x/4 API is a mess. a hack.

Very often there is a mix of modifying the data structure and the look of a form element with the same method call. For instance, putting text in a table element forces the appending of a ':' character to the text. To get rid of this you must make your own UI element routines.

Memory seems to be divided into 'Chunks'. There are two kinds of references to chunks, MemHandles and MemPointers. Both must be locked before they can be used. MemPointers cant be moved in RAM, MemHandles can. Call the method to lock a memhandle and you get a mempointer.
want to unlock a memhandle? you cant use the mempointer, you need the original memhandle.

You have a Form, filled with user interface elements. In the UI definition file you give each element an ID. FrmHideObject will hide an UI element from view. What do you pass to FrmHideObject? The element ID? no! the element Index. How do you know an element's index? You call FrmGetObjectIndex -which by the way wants to know what form you're working with even though UI element IDs appear to be unique across all forms. How does an Index relate to an ID? it doesnt!

Some, maybe a lot, of the frustration I've had with PalmOS is the limitations of c. PalmOS is largely responsible for this because c is the only language the actively promote. For instance to put some debugging info on display in java (Regular Edition) its System.out.println("("i="+i+" message of "+message"; With IBM's VAME they have the mother-of-all classes to wrapp the whole PalmOS API in one class: Os.FrmCustomAlert("i="+i+" message of "+message, null, null). In c its char msgbuffer[80]; StrPrintF(msgbuffer, "i= %d message of %s", i, message); FrmCustomAlert(FatalErrorAlertForm, msgbuffer, null, null); (that may not compile exactly but its close enough to show my point). (note that FatalErrorAlertForm being a form and not a UI element doesnt need the ID to Index lookup)

I have a pulldown list that I want to populate with records from a database. DmGetRecord and funky pointer casting I am able to pull out the records. Now I need to use LstSetListchoices. On Page 259 and 260 of Palm OS Programming 2nd Ed from OReilly, the have pointer arithmetic that rivals a space launch to get a two dimension array allocated and deallocated in the right way, at the right time. Java's collection classes such as java.util.vector would make an operation like this as simple as putting things in a bag (eg. listChoices.add(element)) and handing the bag to the UI element (eg. java.awt.pulldown.setElements(listChoices);)

The last straw was today, under POSE. I had my form up, I had a pulldown list on the screen, I put up an Alert window. POSE reports that this third layer of UI is too much. "Windows cannot be under forms because they can't be redrawn". It appears there is some limit of two layers of UI elements on the screen at one time. Im not working with c and PalmOS 4 anymore.

I have been reading about PalmOS 5 this week and the new features are very exciting. A bluetooth API, plugable audio/video codecs, a sound API, etc.. I am writting this letter to say please make a rewrite of the most basic APIs the priority for the next OS release, or at least wrap the whole thing in a usable, object oriented framework. Metrowerk's project along these lines is encouraging, but PalmOS prides itself on its easy of entry for software developers. This is only possible with free tools. Please promote Java on the Palm platform and include a JVM and Java class libraries on every new palm. Promote open source languages such as Python (see the pippy project), ruby, and perl. Put PalmOS developers 20 years ahead of where it is now by unshackling developers from the "portable assembly language" that is c and showing them a threaded, object oriented, garbage collected environment with Model View Controller user interface elements and remote method calls and mobile code (serializable) and everything else that makes java wonderful!

Thank You
Don Park
http://www.beehivegames.com

It is not about replacing 3rd party apps. DB4 and AN are both handcuffed by core os limits. I own both products and am on their lists. People request more categories, or multiple categories per contact, more address fields ect, the answer is ALWAYS that the are limited by the constraints of the OS. If they add such features they lose compatability with the desktop.

Outlook has this ability and this is what we want on our devices. This is not a Desktop in our pocket request. These are basic PIM funtions, what Palm does best.

I love having choice and do not wish to duplicate the functions that 3rd parties provide. You are right about the fact that they can do this better than you can. That is their role. But without Palmsource making core changes 3rd party software cannot fix what is missing to begin with. This is your role.

Please do not take this as bashing. I only wish to add constructive criticism. I am very excited about what is to come. I love your product and only hope to be helpful in improving it. I hope for an exciting and long future for your company. But please do not underestimate the value to the end user of such a change.

Thanks Mike again for sharing with us a glimpse of what is to come.

Ed:
If you do speak to Mike again please ask specifically whether the core app change here described are in the works.

You know what? I think I can trust those guys.

Let's put it in perspective: They have produced some excellent products over the years. Sure they made some mistakes recently, but they are trying to put things right. Their efforts to recover seem to be genuine and they are clearly addressing the issues in a forthright and sensible manner.

Now that we have some idea of what to expect, 6-9 months doesn't seem so long to wait, in the scheme of things. Like everyone else, I'd prefer 6-9 weeks (or days). However, I understand how important it is to get these things right (the first time). There is no sense in rushing to release some 'half-baked' code, just for the sake of getting something out.

Let's give them a chance and see what they can come up with.

(Ed, this is like my third post on this thread - don't shoot - I mean well.)

I've read the interview and comments here (as of Sun, 2/10, 3PM ET). (I am Graffiti-ing this Comment on my CLIE for later HS C&P.)

Like many, I am disappointed that OS 5 is just a wee more than OS 4-for-ARM.

However, I can understand the strategy Nagel is using. Better "something" now rather than "everything" later (or much later). Or not at all -- viz, Copeland, Taligent, or Nokia's Palm-GUI-on-EPOC.

Several points --

1) I hope PalmSource will take a true leadership role. Hearing that a replacement for DOC is in the works would be a good sign. Let's not leave everything to third parties. If PalmSource wants third parties involved, issue a Paper stating an objective and invite feedback from developers.

2) It's not good news hearing that third-party developers are not financially flourishing. But that PalmSource exec is 1,000% wrong when he said (quoted at news.com), "How do we get the painter...the taxi driver...and the executive [to buy PalmOS apps]? We can't go to CompUSA." If this is his thinking, he should be relieved of that jurisdiction -- fast. PalmOS developers will go bankrupt under such a limited vision.

3) For PIC readers who say they don't want a "handheld desktop replacement" -- it's too late. You already do. PalmOS devices are virtually equivalent to mid-1980s Macs.

4) I don't know why anyone needed to go to the PIC article 20x to look at the prototype CLIE pics. I grabbed photos off the Net and put them in my CLIE to view at any time. Shame, not using your own Palm device like this! (AlbumToGo is free! And PictureGear Pocket in built-into CLIEs. And, yes, the photos are fine on my 160x160 mono CLIE.)

5) The separation of the OS from the hardware is beginning to make for schizophrenic reading. Did "PalmHard" present at the Conference? Do they have a press liason? Is there someone to explain their idiotic decision not to make the "Universal Connector" truly universal? Why can't they license it? I hope PalmOS licensees will band together to create their own UC. It'd serve "PalmHard" right for bringing PPC's Connector Babylon to PalmOS!

People want better PIM stuff, because the PIM stuff is the main reason most people bother with Palms or PocketPCs. This is a true thing. People also want better hardware so they can do more things and be more free from the shackles of the desktop and the laptop. Also true. People want a better underlying OS, to overcome limitations, and bugs, and what all else is plaguing them. True. People don't want Palm to become monopolistic and eliminate most third parties. True again. People want new things that they never thought possible, enabling them to do new things, or old things much much better. Almost Always True.

But most of all...

DON'T BREAK ALL THE STUFF THAT MADE ME LIKE MY PALMOS DEVICE IN THE FIRST PLACE.

Can you IMAGINE how hard it is to balance all these people's desires? What order do you go about them? You almost have to do them all eventually, or people will turn to better things. The previous directors of Palm did not know how to go about these things, so they stuck with what worked, and shackled the entire "Palm Economy" to the past. Slowly, and surely, the new entities are working to allow us to step into the future by, first and foremost, unshackling the hardware from the software. Now they're building a stable foundation, then they will see the potential of their foundation and build great things on top of it.

But they are doing it slowly, taking a wait and see approach, and *listening* more than they ever have, to their community. This may be too slow, or too late, but Palm, PalmSource, and the licensees are NOT dead yet, and in a year or two, we may all have something in our hands that will wipe of every smug PocketPC users smile, and be able to say "oh, gee, you paid $100 more for THAT?!!"

It will not do too much, it will not do too little. It will make me happy, and I'm an idealistic Palm dreamer.

Jahmal
What's WRONG with the Crusty Edge of Innovation??!!