On Fri, 2004-11-26 at 22:05 -0500, Bryan J. Smith wrote:
> On Fri, 2004-11-26 at 18:19 -0500, Mike Dittmeier wrote:
> > I chose dovecot over cyrus because of the documented install steps that
> > I found for cyrus. The fact that I would have to create imap accounts to
> > be used with cyrus v.s. using the list of existing unix users as the
> > list of email users. It was just a simple install for someone who had
> > not installed or configured either of the imap servers before.
>
> Dovecot and Cyrus IMAP servers were really created for different
> purposes. A big one is real users for the former, virtual users for the
> latter.
>
> Dovecot is nice, although I found the version included with Fedora Core
> 2 (CL4.0) to be a bit buggy (not unexpected for a ".0" release). As
> such, I modified the old SPEC file for UW-IMAP 2001 and rebuilt packages
> of UW-IMAP 2004a for Fedora Core 2.
>
> > I do intend to visit the mater further as far as security, stability,
> > add-on features, and speed.
>
> Cyrus IMAP is good to learn if you have virtual users, need ACLs, etc...
> Especially if you deploy something like Bynari InsightClient or another
> collaboration client that uses an IMAP-store.
>
> > The idea in mind is to create a simple install pack that will reduce and
> > standardize the installation of smb servers for customers as well as
> > spreading the work on Linux.
>
> You mean something like SME Server (fka E-Smith Server)?
> http://contribs.org/modules/news/
>
> It's based on Red Hat CL3 (Red Hat Linux 9), and starting to age a bit
> now. Earlier releases date back to the Red Hat Linux 6 timeframe. I
> wouldn't use it for a gateway (something like IPCop is much better), but
> as a web-administered LAN server, it works most excellently.
>
> Development has been stagnant. Some of the contributors I personally
> know were interested in updating to Fedora, and I started to get
> involved earlier this year. But a lot of general talk about moving to
> Debian seemed to derail a lot of interest -- especially when Fedora is
> not only the direct route, but most of the Debian suggestions seemingly
> came from people not interested in doing a lot of the leg work.
There is a large difference between the project you stated and what I
have envisioned. My plan is to produce a package of templated
configuration files that will allow for a fully functional server system
froom the get go, and simple step by step instructions to modify needed
areas. Examples of areas could be as simple as changing the dns domain
name to the ip subnet the network runs on. The idea is not to create a
white box or black box solution, but to provide a replacement solution
to higher cost proprietary systems.
>
> > A large portion of the business industry is small mom and pop
> > businesses, that I feel have been left out of the linux world.
>
> Most of it is due to unrealistic expectations. If you're a consultant
> planning to cater to mom & pop shops, I don't want to deter you, but
> it's really not an area where you want to be charging money. I find it
> has the same problem as many other industries, the savings are expected
> across the board, including how much people expect you to charge them.
I agree to an extent. I have tailored towards smaller shops for 4 years
and have a very nice business from it. The approach to pricing is one of
your keys. Learning to stop putting out fires is the other. If you can
provide a preventive maintenance service that is less costly to the
customer than emergency services, or even pre paid at a fixed rate, you
will have more customers than you know what to do with. I have for 2
years had to turn away many customers because I have been unable to find
reliable staff, that can show up to work on time or stay till the job is
finished.
>
> I often got stuck trying to solve unrealistic application requirements,
> where the shop got themselves into using a mission-critical application
> that was very eccentric.
This is common place even in the windows world. But the truth is most
small businesses need very few mission critical application past a good
accounting package. But you open yourself up to a larger market if you
can provide a custom application to solve their problems. I have seen
the medical industry suffer with unstable windows deployments of medical
manager, and many other commercial applications. The same is true of the
insurance industry. Both come from the diversity of the different
practices and areas of expertise. Pediatricians have different records
needs than internal medicine or even physical therapist. Commercial
insurance companies have different application requirements from home
owner insurance. If you develop apps for these different lines of
business on a stable platform, to begin with at a substantial lower
cost, most businesses are willing to divert the savings to development
cost. Putting those software projects on sites like source forge and
getting the world to help you develop a stable app that works alieves
some of the stress and cost of development.
>
> > Commercial software developers don't provide as many aps for linux as
> > they do for Windows for a reason.
>
> Marketshare? But even beyond that, there's another reason. Because
> Linux _is_ more popular than we reality.
>
> When you see Windows, you know it is Windows. Microsoft ensures this.
> They don't cater to vendor black boxes -- at least not until recently.
> And even then, it's still the "Microsoft support is external" issue for
> an OEM/vertical vendor.
don't we become the oem vendors when we deploy linux solutions?
>
> That's why there are literally millions upon millions of vendor black
> boxes out there running Linux, but no one knows it. Because the reality
> is why would a vendor retail/vertical a Linux application when they can
> sell you a whole box they support? Something they can support from
> hardware to end-user app.
HP, IBM, and even Dell are selling and supporting linux installs more
and more every day. I think it's a worthwhile investment.
>
> And that's where Linux is succeeding brilliantly. As a more reliable,
> hardware-to-application, black-box platform. There is little to no
> incentive to sell an end-user installed Linux application atop of
> another platform. Because the OEM/vertical vendor can control the Linux
> platform itself, unlike Windows.
>
> > That can be changed very quickly if we can convert a large portion of
> > these smb businesses to linux. My plan is to pull together alist of
> > commonly used packages that provide all of the services and apps used
> > by smb business to include accounting packages, pos sytems, and so on.
>
> I wish you the best of luck. If it's a Freedomware project largely done
> for free, it will do well.
>
> But I have found it's difficult to consult for hourly rates for mom &
> pop shops. It's far better to sell an "application-specific black box"
> solution that leverages Linux, something you can control from not only
> just the end-user app, but all the way down to the OS and hardware too.
> Something you can't with a Windows solution.
>
> And then you can sell that solution, possibly with or without SLAs
> (depending on the applications), at a price-point and TCO far lower than
> a Windows solution. And it caters far better to the customer too.
>
> These are just my personal observations and experience. I personally do
> not consider myself a shrewd enough businessman to make a good buck, so
> I might not be the best person to listen to. With that said, if anyone
> needs a good system integrator to design, cater and support such total
> black box solutions for network border, network security, file services,
> development/release control, etc..., I'd love to find a partner.
>
>
-- Mike Dittmeier, RHCE, MCSE mike@bluecrabtech.com----------------------------------------------------------------------- This list is provided as an unmoderated internet service by Networked Knowledge Systems (NKS). Views and opinions expressed in messages posted are those of the author and do not necessarily reflect the official policy or position of NKS or any of its employees.
This archive was generated by hypermail 2.1.3 : Fri Aug 01 2014 - 18:34:41 EDT