Let's hear it from the horse's mouth.

[Mark]

Please udnerstand i'm not trying to be combative or argumentative. But, a bunch of people are going to be wondering. I, too, would like some clarifications..

Are all the features (that are physically possible, ie, platform capabilities) going to be shared between V3 and the new platform?

Are we really going to have QOS, and a few other important new features? I realize you've made general statements previously... but it would be nice to have an enumeration of the "for sure" things and some of the "expected". Maybe even some "never gonna happen" things, too.

[lonnie]

I can confirm that VxWorks releases will be free just like the StarOS Linux releases. We hate to be required to keep paying for a product we have purchased. Yearly maintenance means the guys have captured as much market as they can and growth has stopped, so they must milk the current users. R&D and support should be factored into the revenue from selling the product.

I feel that a license should never expire. It might not continue to have new features added and you might have to pay an upgrade fee for really new features, but never to simply use the features you had been accustomed to using.

We are treating the VxWorks as a special release since each copy will cost us a royalty to Wind River and we need to control the deployment to be able to account for them and pay that royalty.

Also, VxWorks does not function well on any old hardware. It can be made to run on almost anything but that image will then only work well on the type of hardware. This has meant it cannot be used for our main software only product, since people expect to be able to pick and choose their own hardware. This means that v3 will be general purpose which means it cannot be tuned to the hardware in the same way we can do with VxWorks. Simply put that means the VxWorks and the hardware we use will always outperform an equivalent CPU speed on the v3 release.

It is not an evil plot to get rid of v3, it is simply the way it is.

v3 will eclipse v2 in terms of features. It will have a similar system design to the Vx release. The overall structure is geared to be able to expand the feature list and not have limits that came from the very first units we developed. In order to do that we will be breaking with the ability to upgrade v2 to v3 in place. It will require that you change the flash, preferably on the ground. The OS is different and the config layout is not even close, so it will be nearly impossible to have Linux v2 upgrade itself to FreeBSD v3 and save all settings. We might give it a try but I can see so many potential troubles.

QOS, advanced bandwidth control, advanced Atheros driver, better control of Hotspot, PPPoE Server and client, on system radius using the system flash, Windows GUI, ....... etc.

People need to check the release history and just for fun compile a list of features we added to v2 AFTER we said no new features. We can't help ourselves. It is in our blood to keep adding things that make life easy.

Now I can say anything I want and you can challenge what I say. In the end it all comes down to faith and trust that we will keep adding those features. To my mind, the best indicator of that is the very history of our software. It has truly expanded from some very basic roots. I am not sure if anybody still remembers our very first picoBSD router release, but that system was very simple, yet it got the job done. In truth we still have a user in our local network who is using it. For a simple client it was great and it was solid. The solid part is what we focus on. Once it is solid we add the fancy.

We have always done it this way. I see little reason to change. Do you?

[Mark]

PLEASE try to make some kind of v2-v3 upgrade mechanism. Even if it doesn't maintain the settings, I really, REALLY need to be able to move a lot of CPE to V3 in the future. Takeing these down from rooftops is not a viable option.

Even if it means I would have to upload a tiny system that reformats and boots to a simple install that then permits a new firmware upload via a default IP and then I reconfigure.

[ninedd]

I can confirm that VxWorks releases will be free just like the StarOS Linux releases. We hate to be required to keep paying for a product we have purchased. Yearly maintenance means the guys have captured as much market as they can and growth has stopped, so they must milk the current users.Personally, I don't agree. :) I'd rather have a small upgrade fee (say $10) on every major or even semi-major release so that there was a constant, built-in incentive for you to develop & support new features as quickly as possible. I know this is a catch-22. You want to make sure that there are no barriers for everyone to use the latest release, but there are also time you (Lonnie) say things like "I don't have the time to...." and I think to myself, "that's because you have a product that's $40 for life, with free lifetime support". Frankly, I think that's too good of a deal. :)

As customers, I also don't like paying 'license fees' that never end for no good reason, but on the other hand, I'd glady pay less for a CPE version and more for an AP version, and I'd glady pay a yearly service contract or license fee for the AP version, since that's what requires the support questions. I'd also be glad to pay a small amount, like $10, every time there is a major version. If I choose not to upgrade, then my old version AP is EOL and is supported for a limited time - like 6 month's after a major version is released. That wouldn't mean we had to pay anything extra - we could continue to use it forever for free, just no free upgrades after a year. As customers, I don't think we can ask or expect to have a $40 product that is free forever, when you have ongoing costs for support and development. :)

I know there is also a complexity and a cost involved with tracking all this stuff, so it would have to be as simple as possible. If I was in your shoes, I'd do something like this.... Charge $20 for a CPE version, $50 for an AP version and AP's would be $10 to upgrade to the next version. If anyone balks at $10 to upgrade thier AP's from ver 2 to ver 3, they're way to cheap. :)

[ninedd]

PLEASE try to make some kind of v2-v3 upgrade mechanism. Even if it doesn't maintain the settings... Even if it means I would have to upload a tiny system that reformats and boots to a simple install that then permits a new firmware upload via a default IP and then I reconfigureI agree. Plus there are tower top AP's that I'd rather not climb too if I can avoid it.

[lonnie]

I suppose we can try and make an upgrade version that writes the new image and has default settings to allow local access.

We'll think about charging you guys more if you really want it. Just joking, of course.

Wouldn't you worry about us stalling fixes so we could release an upgrade with a fix, thus forcing you to buy the upgrade for the fix?

[Mark]

I suppose we can try and make an upgrade version that writes the new image and has default settings to allow local access.

We'll think about charging you guys more if you really want it. Just joking, of course.

Wouldn't you worry about us stalling fixes so we could release an upgrade with a fix, thus forcing you to buy the upgrade for the fix?

Please don't charge for upgrades.... but I am and have been buying steadily, and plan on having over 500 star-os licenses by the end of '08. Most have been bought through US distributors, due to shipping pains for small amounts from Canada (bought with hardware).

The way I see it, your path to wealth here is a VIABLE INEXPENSIVE CPE. WISP's buying AP licenses is nothing... compared to WISP's buying CPE licenses.

Your potential sales when we can count on a REASONABLY PRICED STAROS BASED CPE (the WRAP is real close, it's about 15% high in price, and lacks a good affordable enclosure... plus, it needs to be FCC certified...) will go up by a multiple of 20, from current. And, that with virtually NO more support time. on your part.

[go.fast]

I do not think I can afford to pay for incrimental upgrades.
Major upgrades possibly, but the typical upgrades would price a lot of people out of the market.

I would also like to suggest that you should do what is needed to hire an extra coder to help you move forward faster.

Not sure how many of you there is hacking away, but one thing I noticed is that a common complaint is "time".

Time is our biggest competition, and I'm not talking about the time it takes for you guys to get things done.
Just time in general is what makes us wisps lag behind.

An example not related to you guys is the 900MHz, it too Trango a long time to come out with their 3 meg 900 after announcing they would.
When I put my order in for my first dozen cpe's the cable company turned on the subdivision that I was planning on serving and scoffed up half of my prospective subs.

And then the time it took for them to work out the bugs that they were having made me hold up on buying more and the cable company scoffed up even more potential subs.

So time is the competition not the cable company.

You guys need to hire another couple of coders to code a "time machine"

George

[lonnie]

The tough thing right now is the rewrite. It is all new code and little things are being found under intensive testing. Once we get a piece of code fixed it becomes a solid foundation to keep building upon.

Time is a problem. I'm not sure how many of you have ever been in a large software project, and this is a large software project, but in large software projects the general rule is that if you want to delay the project you simply hire some more programmers.

There is no clear reason except that more people means more discussion, which leads to unproductive time and it means that new guys do not understand the old code and thus add hacks instead of fixes. I suspect it would take a junior programmer 2 years to come up to speed with what we are doing and a senior programmer would take 5 years, and might not ever be useful. The reason for that is that a senior programmer already has their bad habits well learned and their style is ingrained and will most likely not mesh with the other team members. It is constant conflict and jockeying for power. Not productive at all.

So, with years of experience behind me, I have decided that we continue as we are. We will be too late for some projects, but the current software and WRAP boards works pretty well and keeps us in the game. It is the new stuff that we are looking to for the ability to add new things. And that was the goal when we first set out on v3. It took a turn with the Vx, but the v3 will reuse most of the Vx code, so in reality v3 is being built. You just can't see anything but the foundation yet.

[Mark]

I would tend to agree... A programmer without experience would slow you down, simply due to the fact he won't be very productive, and you'll have to fix his errors. An experienced programmer hasn't learned YOUR habits (be they good or bad) and will also cause issues.

However, time is a fungible commodity. In other words, there ARE things you could have others do - testing, some support time, etc. For instance... You're talking about adding a list serve. This WILL take time.

Catalog all those things that take your time - those that enhance your forward motion, those that detract, and those that are priority - such as family.

I would suggest, for instance, someone to run this board for you, someone to run the lists, perhaps some of your testing. There is wisdom in adding more team hours in a day... Just use them carefully.

[ninedd]

Wouldn't you worry about us stalling fixes so we could release an upgrade with a fix, thus forcing you to buy the upgrade for the fix?Well, that's a good point. Certainly I've felt that way with US-Robotics. On the MP/8 I-Modems, they released V.90 that was broken, charged for Firmware updates and then EOL'd the product without it ever working. We were even the ones that were able to demonstrate to them exactly how to duplicate the bug, and have a thank you letter to that effect from them. They still wanted huge $ for a bugfix, and yes, I felt like a hostage.

However, with you (maybe it' Stockholm Syndrome??) it would be different. If upgrades were a low cost, like $10, then that wouldn't be an issue. You'd

I'm just saying, if you have 1000 AP's in the wild, and it's $10 each to upgrade an AP from ver 2.17 to 3.00, then that's $10,000 to you, and it's really insignificant to each of us. We're all going to spend many hours upgrading and learning the new features, +-$10 is not a make it or break it deal for any of us. However, the $10,000 is quite significant.

Then, when people start asking for the wipeyerbutt routing protocal to be added - something that is not a 3.01 fix, but a new feature that qualifies as a 3.10 release, if that's $10,000 more for you, it would have to get a higher priority than a free fix because maybe it'll expand your market $40 at a time. :) CPE licenses are a different thing, since an WISP could have 100's, so an upgrade cost could be prohibative. Of course, a CPE license only really has to have Infrastructure mode, Bandwidth Control and a routing for a /24 at most, so there should be less to be broken and need fixing.

I started thinking this when you were not doing Atheros support way back when. I think you're stumbling block was $20 or $30K for their kit. I thought to myself, if it's $20 upgrade x 1000 AP's, that's a significant new feature, that wasn't in the product when I bought it originally, so I have no expectation that you should add it for free. Same as QOS or other features that don't exist right now - when they are added as new features, I see no issue with charging a small amount for that upgrade. :) Bug fixes should be free, but new features should be charged for so that you want to mature them to market quickly.

Now that everyone hates me.... :)

Todd

[ninedd]

Please don't charge for upgrades.... but I am and have been buying steadily, and plan on having over 500 star-os licenses For CPE licenses, I agree. It's the AP licenses where the new features are added. It's also the potentially large CPE base that is potentially very profitable for you, and potentially very costly to upgrade for a WISP, so I'm not suggesting charging for upgrading a CPE version. As a matter of fact, I was suggesting that a CPE version should only have Infrastructure mode, bandwidth control, /24 routing and that's about it - and that the CPE versions should be even cheaper. I'd like to see $20-$25 for stripped down CPE version, and $50 for an AP version, with $10 for upgrades (not bug fixes, but feature upgrades).

And, that with virtually NO more support time. on your part.Also correct. That's why I'm suggesting to charge just for AP Upgrades. Not only should a CPE license have limited features, but when you teach the WISP how to do bandwidth control on the AP, then he knows how to do it on each of his 200 CPE's anyway, so you don't need to charge 200x for that support.

[ninedd]

I do not think I can afford to pay for incrimental upgrades.
Major upgrades possibly, but the typical upgrades would price a lot of people out of the market. I would also like to suggest that you should do what is needed to hire an extra coder to help you move forward faster.... GeorgeExactly my point. I don't want to see $10 every time you realease a beta or a bug fix or a maintenance upgrade, and I don't want to see $10 for every CPE a WISP has, but it could help solve the ''time'' issue to have more programmers on the job, and those programmers cost money. If we can have a $10 upgrade in a month, or a free upgrade in a year, I truly think that everyone would gladly pay $10 for each AP.

I don't know what the average / WISP is... 5 AP's? 20 AP's? 50 AP's? If you have 5 AP's, then $50 upgrade them all to new features is not a signifigant "deal-breaker" cost for a WISP. Even if you have 50 AP's (and therefor 1000+ subs) then $500 to upgrade them all is still relativly insignificant for that business.

Either way, go.fast is correct... Our gut feeling is that we can't afford another cost, but in reality, what we can't afford is to wait for the competition or the industry itself to pass us by. :)

- Todd

[Mark]

For CPE licenses, I agree. It's the AP licenses where the new features are added. It's also the potentially large CPE base that is potentially very profitable for you, and potentially very costly to upgrade for a WISP, so I'm not suggesting charging for upgrading a CPE version. As a matter of fact, I was suggesting that a CPE version should only have Infrastructure mode, bandwidth control, /24 routing and that's about it - and that the CPE versions should be even cheaper. I'd like to see $20-$25 for stripped down CPE version, and $50 for an AP version, with $10 for upgrades (not bug fixes, but feature upgrades).

Also correct. That's why I'm suggesting to charge just for AP Upgrades. Not only should a CPE license have limited features, but when you teach the WISP how to do bandwidth control on the AP, then he knows how to do it on each of his 200 CPE's anyway, so you don't need to charge 200x for that support.

I use the router version for CPE. I think the router version could not miss much and be complete.

I need QOS, NAT, Routing, layer7 filtering, CBQ, dhcp server, pptp server, and a pile of other features on my CPE. I don't mind the $40 price tag for that. I would severely mind that being $100.

[go.fast]

I don't mind the $40 price tag for that. I would severely mind that being $100.

I agree, 100.00 is not an option, 40 is do-able but less is better.

What it all comes down to is quantity.
If enough units are sold, then they should consider a lower price.

But I don't know what enough is.

George

[ninedd]

I need QOS, NAT, Routing, layer7 filtering, CBQ, dhcp server, pptp server, and a pile of other features on my CPE. I don't mind the $40 price tag for that. I would severely mind that being $100.$100? No one suggested $100 I don't think - unless I have a typo? Me confused...

Really, I probably got way to detailed and long winded and really was thinking out loud. Obviously, without even knowing how many AP's Valemount has sold out there in the wild, I really can't sit here in my underware and re-design their marketing plans. :)

My only points were 1) I think it's too much to expect to get a great product with free lifetime support and fee lifetime upgrades for $40, and b) that a small fee for upgrading to new features into the AP version would not impact us each very much, yet could be a signifigant incentive to allow faster feature releases.

That's my story, and I'm stick'n to it.

Todd

[Mark]

$100? No one suggested $100 I don't think - unless I have a typo? Me confused...

Really, I probably got way to detailed and long winded and really was thinking out loud. Obviously, without even knowing how many AP's Valemount has sold out there in the wild, I really can't sit here in my underware and re-design their marketing plans. :)

My only points were 1) I think it's too much to expect to get a great product with free lifetime support and fee lifetime upgrades for $40, and b) that a small fee for upgrading to new features into the AP version would not impact us each very much, yet could be a signifigant incentive to allow faster feature releases.

That's my story, and I'm stick'n to it.

Todd

I don't think anyone suggested $100. I guess the point I was trying to make, that in my view the CPE needs to be just about as complete as the AP, and that I'd rather not see a different version for CPE. I have CPE, btw, that are also AP's with a 2nd radio in it, where I relayed to a 2nd customer out of sight.

If the versions become split between CPE and AP, with the only differnce being AP functions disabled for CPE with a cheaper CPE license, then I see it as worthy, perhaps.

I really don't see the issue, however. As far as Valemount goes, thier future could be dramatically improved by finding a way to make STar-OS based CPE competitive. As I said, I believe it close... as small quantity CPE of any kind of worthwhile quality vary from 180 to 450 dollars. WRAP based CPE can be b uilt for 210 to 275 dollars, depending on choices and vendors, It just needs that impetus to get "the ball rolling", and push it into mainstream - and really, more than price, it needs to be someone to integrate it neatly and affordably.

Ascendance does, but raises the price too high.

We make our own, we add about 25 dollars to a WRAP kit for enclosure, labelling, ethernet moisture sealing, and mounting - and the worst priced part of it is the motherboard standoffs!

[go.fast]

and the worst priced part of it is the motherboard standoffs!

I think I pay less than a quarter per standoff.....

George

[Mark]

I think I pay less than a quarter per standoff.....

George

So what do YOU use?

I need something I c an epoxy or self-glues to a polycarbonate surface. And about 3/4 inch off the surface.

[go.fast]

So what do YOU use?

I need something I c an epoxy or self-glues to a polycarbonate surface. And about 3/4 inch off the surface.

https://secure.microplastics.com/detail.asp?part=
pcbsupportonadhbase&fam=cbhardware