SUSHI - Software Update Service (Hyperlinked, Interactive)

[wardmundy]

We've got a lot of hobbyists and tinkerers. We're trying to balance between the business users and this group. My vote goes for starting at a low annual fee and evaluate things again in about six months. If the revenue isn't meeting the needs of the project, we can adjust it. In that way, at least the early supporters get a financial break.

We're also reconsidering fixed grants to other projects. We want to contribute, but I think the consensus is that annual grants to worthy projects with no predefined entitlements makes more sense. Then we'll be in a better position to balance the needs of our project against support of future development in other areas.

 

[riccardocapra]

I really appreciate two things from you: you're really onest and you say thae things as they are. I'll sure pay 30$ for server, as I donated to the project. Tell me how.

 

[jucha]

Excelente. Apoyo esta idea.

 

[g711]

I feel that this overall would fill the needs to start with and does not exclude anybody. But what about a 2nd tier level for those in the commercial arena where there could be a list of applications with a price tag attached to get the project done and once done either released to the public or licensed via a USB key purchased from you to further fund other projects, put food on the table etc. I for one would love to see a nice GUI with pretty pictures, graphs and drill down stats for ACD and our own DUH (read backwards) that does not take up so much real estate on the monitor etc. etc.

 

[mkhurrum]

I agree with Ward. Lets start a low annual fee and as business users grow - add more services and charge for those.

MK

 

[undrhil]

I agree with Ward. Lets start a low annual fee and as business users grow - add more services and charge for those.

MK

No. I do not like the idea of "Let's charge a *starting* fee now and then we'll increase it later." This always ticks me off when I see places offering a service that will "always be free" or "never go up in price" and then they make good on that by forking the service to add more features under a different name and for (more) money.

Whatever price they pick now, they should stick with. That's why I think they should consider doing it as a monthly fee or maybe more per year than $30. I think most of us are willing to pay more than $30 a year for the update service. The convenience of not having to reconfigure the system from scratch is worth more than $30 a year, at any rate.

 

[jamaica007]

I concur. I would gladly pay the 30 Dollars to help small fee for such great work and development. let me know where to sign up

 

[krzykat]

I would be more than happy to give back some for all the hard work that people doing open source do.

 

[cvegas]

I would be happy to participate at the $30.00 range

 

[bkiesz]

I don't think $30 is bad at all for what we're getting. I would like to caution the powers that be not to get too "Ala carte" with the addons or you'll eventually become what you're trying to replace. But I think Ward and the boys already know this.

For those who are making a business out of this, every addon piece will take away from their bottom line. I know that they can pass this on to the customer but it's pretty cutthroat out there and you don't want to nickel and dime them.

How about this...

$30/yr per system for automated update scripts no matter what... This can be passed down to customer pretty easy. They understand this..

For the SMB that wants the PnP drivers for XYZ card(s), we pay, $50.00... for someone who is in it for the PBX business, they pay a "Site License" flat fee of $50-75/yr for the special driver scripts , and so on...

Bottom line... the guys doing 1 and 2 systems, let them pay for the time saved. for the system builders, let them pay a token fee for these scripts, cuz the more systems out there, the more $30/yr script fees .

My $.02 cents....

Barry

P.S. I have to say that this Distro of a PBX is just awesome and I've loved the nervittles artices from day one and check the site 3-4 times a week looking for new ones!

 

[wardmundy]

P.S. I have to say that this Distro of a PBX is just awesome and I've loved the nervittles artices from day one and check the site 3-4 times a week looking for new ones!

Thanks. And some day, I hope I'll have time to write another article or two.

 

[Jay Levitt]

Forget about the fee. Here's what I don't get: Why are you inventing a whole new layer of scripts that (at least so far) can't tell if they're re-applying the same changes, can't tell what's already been changed (by previous scripts or by the site admin), and are apparently so fragile that they have to warn me against touching the keyboard?

You're based on CentOS, and that's based on rpm/yum - which, while far from perfect, handles all this stuff easily. Why isn't PIAF yum-based as well? I searched before posting, but "yum" and "rpm" are three letters, and your forum software prevents searching for such short words.

 

[kenn10]

Jay, if you've been around this forum for any period of time and read the postings, you'd know that you can search like this:
"PIAF:yum" to get around the 3 character limit.

Secondly, you're coming on way too much like a basher on these guys who put stuff out for free on their own time. Do you have something to add to assist with the process or are you just posting to stir up a flame? Looking at your posts, you've had little or nothing good to say about anything to do with PIAF. Perhaps Trixbox might be better suited for your needs and level of experience.

Confucius say: `The superior man is satisfied and composed; the mean man is always full of distress.'

 

[thunderheart]

Does kinda come off ...

... well, boorish. Not to mention I don't think this is the right thread for it.

Jay, people generally react better to a more constructive tone. Ward et. al. certainly must have chosen their design philosophy after a good bit of thought. They after all aren't a bunch of newbies who just stumbled on Asterisk, got jacked on Red Bull, and threw together a distribution. They are guys who have been around the community (as far as I can tell) for quite some time, having made contributions to other distributions that are signifigant. They are known and well respected.

They sure seem willing to answer questions when asked in a civil manner. You might give that a try if you really want answers.

Dallas

 

[Jay Levitt]

Sorry; I didn't mean to come off as boorish. In my defense, I live in Boston, and there's something in the water here.

Let me try again:

I am new to Asterisk. I've found the Nerd Vittles posts and distributions very helpful, but I find PIAF's extra layer of shell scripts to be awkward. In fact, I'm working through the O'Reilly book, and I've found it easier to just build the source from scratch, because I don't know what the scripts do, or where they put things.

I know many people here are upset about the fee; I'm not. I, personally, would LOVE to pay $30 to get a reliable, self-maintaining Asterisk distro. I think PIAF is a great step toward that. But I really wish it would use a more robust packaging system. Is there a reason it's not done with RPMs? The shell scripts are very brittle.

As you can see by my post count, I'm new to the forum, and I don't know if this is an old debate. I tried to search for "yum" and "rpm", but vBulletin is set to disallow searches for words shorter than three letters. I also tried the "PIAF:yum" syntax, as posted by the superior man, but it pulled up many posts that did not contain the word "yum", so I am apparently misunderstanding. I did search for "packaging", and didn't find any relevant threads.

If there is a better thread for this, I'd appreciate a pointer in the right direction. I came here from the link in your blog post, and this is, as far as I can tell, the thread where we talk about SUSHI. I'd search for a more appropriate thread, but, well, see the previous paragraph.

 

[thunderheart]

Better Search

Jay,

Lots of folks complain about the 3 letter search limitation of vbulletin. The best way to search for 3 letter terms is from google using a search term that looks like this:

site:www.pbxinaflash.com yum

At least I think that is the syntax. It is certainly mentioned in multiple threads on this forum.

As for the $30 fee for SUSHI, I really haven't seen anybody complain about it. Actually, what SUSHI primarily buys you IS the script layer you mentioned. If you think the scripts are bad, you simply don't pay the $30 dollars and then you have to do any upgrading (like of the Asterisk source for instance) manually. Maybe that would be the way you would go.

You can still even "upgrade" between PIAF releases for free. You just have to do a "scrub" between them i.e., install from a cd that wipes your harddrive first before installing. So lets say you are running PIAF 1.3 and along comes 1.4. You can "upgrade" by downloading the 1.4 ISO, burning a cd and then booting you PBX with it. Of course that will smoke any extensions, routes, customizations etc. that you have on the box, but it is still a viable route.

However, in defense of Tom and the guys, if you follow the process for upgrading (which will eventually cost you $30 a year) then there really aren't many problems. I've done it a number of times on two separate releases (1.1 and 1.2) and never had a problem.

Tom will also tell you "don't upgrade for no reason" which really applies for ANY software in most folks opinion. If there is a security advisory, or a feature you really need, or a bug that affects you and it is addressed by a new version, then you should of course upgrade. Otherwise, why risk it?

Yellow Dog and rpms won't help in that respect by the way. You can frak your system up with yum or manually rpm'ing just as easily as you can with a script. At least I've certainly managed that trick with ease myself a number of times.

Anyway, I'm sure that Tom, Joe or Ward have a good reason for doing things the way they do. To me it does seem a little different for sure, but I don't too much care how it works as long as it does. I long ago stopped worrying so much about how a system works and concern myself mostly with what it produces.

If you are really interested in what Toms scripts do, he does create a fairly detailed log when they run. I would like to see under the hood myself but I understand his reasoning and the system does work.


Dallas

 

[jroper]

Hi

The original reason for using scripts rather than RPMs was the speed of production, and where possible, to compile from source for different architectures

It is quicker to add a few lines to a script rather than to build an RPM. The methodology is to work out what is broken, what you do to fix it, refer to your log to see what commands you ran to fix it, then roll that into a script.

This is evidenced in say update sources, where asterisk is updated. Because it references a link at digium downloads, provided that they update the link, which brings down the latest asterisk source as soon as it is released. This means that you can have a Digium security advisory sorted in a very short space of time. Where an RPM is used, you have to wait until someone rolls up a new RPM.

The most powerful reason is that if you have a problem with a Digium card, you will have no issues getting support from Digium with our Distro. They find it more difficult to support you at all with a Distro which uses home grown Asterisk RPMs.

Joe

 

[tel0p]

my twist

I wouldn't have a problem shelling out $30 bones.

What's kept me coming to nerdvittles.com for so long is looking forward to getting my hands dirty with some new fangled feature . See, I do represent a tinkerer and hobbyist. Been doing Asterisk since way before GUIs (v.6) and just love how Ward writes and makes it easy for all (most). He really has been a hero to many.

Will this change once SUSHI is implemented? Will the tinkerers among us still be able to read-along an article, see how it works and still understand how it works? Or will it be a 'I paid my $30, click, new feature, done'? I'm really more fond of being under the hood and in some source code. Will this still be a reality for me?

Just wondering.

 

[wardmundy]

The Nerd Vittles approach isn't going away. We may give the SUSHI folks a head start once in a while, but we're still firmly committed to the open source, open forum approach to software development. Having said that, we do enjoy eating, too.

 

[Jay Levitt]

The original reason for using scripts rather than RPMs was the speed of production, and where possible, to compile from source for different architectures

Excellent point; RPMs don't seem to deal cleanly with 32 vs. 64-bit architectures.

This is evidenced in say update sources, where asterisk is updated. Because it references a link at digium downloads, provided that they update the link, which brings down the latest asterisk source as soon as it is released. This means that you can have a Digium security advisory sorted in a very short space of time. Where an RPM is used, you have to wait until someone rolls up a new RPM.[/quote]

Another great point, especially when you don't have a large maintainer staff.

Thanks for the detailed answers. I'd still like to see a more robust packaging system, but I now understand the benefits of the current system.