Been there. At the risk of sounding like Bill Clinton, "I feel your pain."
:-)
One of the powers of Hobo, in my opinion, is to allow you to put something
in front of the client as part of the requirements process. In over 25
years in this field I have never had a written requirements specification,
regardless of how detailed, that was sufficient.
Cognitive science and experience has taught us that there is a HUGE amount
of implicit and tacit knowledge that does not get triggered or expressed
until someone actually sees and plays with a process, or "acts it out". It
is amazing how much "muscle memory" we have. I have learned a great deal
about this, ironically, by studying with a professional jazz saxophonist,
who also has the rare talent of being a great teacher. One of his "tricks"
with me, is to pull the sheet music away when I told him I had failed to
memorize a tune for a lesson. I was often surprised about the extent I
could play what I thought I didn't know.
I have seen the same thing after what seemed like exhaustive and seemingly
limitless requirements specification sessions with key users and subject
matter experts. After months of documentation, we would start on first
prototype, and show it to the user. Then all kinds of exceptions and
"miscommunications" come to light. I am sure you have been there. People
just can't express all of the steps they take to do complex tasks that they
have been doing for 30 years in simple sentences. That is why
apprenticeship has been so important for generations.
This is part of a ongoing dialog Tom Locke and I have been having for the
past couple of years. How do you provide a tool that is *an integral
part*of the requirements ("knowledge management") gathering process
(knowledge ->
prototype -> feedback -> correction) that will output something that you
don't need to throw away.
To me that is the Holy Grail for application development tools.
-Owen
On Sun, Feb 28, 2010 at 7:12 AM, Michael Q <[email protected]>wrote:
> Thanks for your replies.
>
> As it's currently configured, it is quite read and write intensive,
> undoubtedly due to my suboptimal coding etc.
>
> My app is actually a Delphi client-server system which has taken me
> over 2 years to develop and its still way from ready for launch,
> although it has been "live" at one client's site for over a year. I'm
> so dispirited by the difficulties that I've had with it, partly due to
> my inexperience at the start, and partly due to lack of clarity about
> the domain and the client's own understanding of their requirements.
> I've learnt a lot, but it's been a nightmare. Now there are some
> performance problems with larger data sets, and my code is such a
> tangle that it's virtually unmaintainable. I hate it. I'm beginning to
> think that the time needed for me to unravel it might be better spent
> starting again with a really efficient tool (from a programming point
> of view) and getting it right, or more nearly right, using agile
> techniques and launching it incrementally as SaaS a la 37 signals.
> Hence my current experimentation with Hobo.
>
> At this point I almost don't care about my target market's potential
> resistance to SaaS. I just want the pain to stop.
>
> Michael
>
>
> On Feb 27, 7:17 pm, Owen <[email protected]> wrote:
> > I can't add much to what Matt has said, except that a few large apps
> > appear to be fine on Heroku. There is a video testimonial from
> > FlightCaster on the site that you may have seen.
> >
> > Barquin will soon do some extensive load testing that we will be glad
> > to share to the group. It would be great for others to share tuning
> > tips as well. We have a well-tested JBoss/JRuby stack for production.
> > This has more to do with enterprise standards that proven performance
> > gains, however.
> >
> > Will your app be geavily write intensive, read intensive, or both?
> >
> > Thanks,
> >
> > Owen
> >
> > On Feb 26, 8:55 pm, Matt Jones <[email protected]> wrote:
> >
> > > On Feb 26, 2010, at 8:12 AM, Michael Q wrote:
> >
> > > > 1. Is it viable to use Hobo to produce a SaaS multi-tenant app with a
> > > > moderate number of concurrent users? I guess I'm hoping for perhaps
> > > > 2,000 - 5,000 concurrent users. (Beyond that, I guess the income
> would
> > > > be sufficient to subcontract a lot of tuning & optimization). Is
> there
> > > > any information about how best to do this?
> >
> > > The key thing to remember is that ultimately, Hobo *is* Rails. So
> > > pretty much anything that will speed up a Rails app (caching, etc)
> > > will do the same for a Hobo app. DRYML views, by default, are somewhat
>
> > > slower than ERB - but the modular design tends to make it easier to
> > > add in fragment caching as needed.
> >
> > > > 2. Is Hobo any more vulnerable to security/hacking than a regular
> > > > Rails app? Can one rely on using only the normal rails lockdown
> > > > techniques?
> >
> > > I can't think of any reason that Hobo would present any additional
> > > security risk. In many cases, it may actually make security a little
> > > easier - the permissions system makes consistently access-controlling
> > > objects much simpler.
> >
> > > > 3. I noticed a comment on a group somewhere that automated testing
> was
> > > > less necessary with HOBO, presumably because a lot of the logic is
> > > > already taken care of. Is this true?
> >
> > > I'd say that's a matter of the individual testing philosophy.
> > > Personally, I don't tend to spend much testing effort on standard CRUD
>
> > > stuff anyways, so there's not much of a difference.
> >
> > > > 4. Which host would you recommend for a mission critical app? Is
> > > > Heroku up to it?
> >
> > > Well, Heroku's marketing says they've got apps that scale pass 100
> > > reqs/s, so I'd assume they're pretty solid. Whether they're the best
> > > choice is dependent on how much support you need, as well as the
> > > particular needs of your app.
> >
> > > Hope this helps,
> >
> > > --Matt Jones
>
> --
> You received this message because you are subscribed to the Google Groups
> "Hobo Users" group.
> To post to this group, send email to [email protected].
> To unsubscribe from this group, send email to
> [email protected]<hobousers%[email protected]>
> .
> For more options, visit this group at
> http://groups.google.com/group/hobousers?hl=en.
>
>
--
Thanks,
- Owen
--
You received this message because you are subscribed to the Google Groups "Hobo
Users" group.
To post to this group, send email to [email protected].
To unsubscribe from this group, send email to
[email protected].
For more options, visit this group at
http://groups.google.com/group/hobousers?hl=en.
