project planning interface thoughts (job?)

[Tim McEwan]

Hi All,

I want to build a MS Project/Omniplan-style project planning interface in a web browser.  I'm really keen for jQuery & HTML5 and stuff, but if it's going to be a ludicrous undertaking, I'd settle for something like Flash.  We basically need to allocate team members to tasks and then time to the team members, but ease of use for non-techies is key, so we need things to be draggable.

I'm thinking me (Rails moderate, no jQuery) + one Rails/jQuery guru for two months @ 35 hours/week oughtta do it.  Am I dreaming?

If you'd be so kind, please send me your thoughts on:

- would you stick with Rails/JS or go with Flash/Air/whatever-that-microsoft-one-is?

- do you think the time estimate is feasible?

- how much is a Rails/jQuery guru of the required calibre?

- would you like to be that guru? *

- no really, am I dreaming?

* It'd have to be onsite at UTS, if you're interested.

Cheers,

-- 
Tim McEwan
Sent with Sparrow

[Julio Cesar Ody]

Maybe you don't need to replicate MS Project or Omniplan's interface
for it to be usable. Why not a Google Calendar-like one? Personal
preference aside, for the task you just describe, it's suitable.

> - would you stick with Rails/JS or go with Flash/Air/whatever-that-microsoft-one-is?

Rails/JS. Though I'm no expert in Flash/Air to judge how easier/harder
it would be. I tend to think you'd be painting yourself against a
corner by going down that route.

> - how much is a Rails/jQuery guru of the required calibre?

If the requirements are really just the ones you've mentioned,
back-end wise it's dead easy to built. As for the interface, not so
much. But perhaps you don't have to start from scratch →
http://www.queness.com/post/656/10-beautiful-jquery-and-mootool-calendar-plugins

> - do you think the time estimate is feasible?

Yes. Should the requirements stay as such.

> --
> You received this message because you are subscribed to the Google Groups
> "Ruby or Rails Oceania" group.
> To post to this group, send email to rails-...@googlegroups.com.
> To unsubscribe from this group, send email to
> rails-oceani...@googlegroups.com.
> For more options, visit this group at
> http://groups.google.com/group/rails-oceania?hl=en.
>

--
http://awesomebydesign.com

[Tim McEwan]

Thanks Julio.

To clarify - this system will be custom-built so it won't work in exactly the same way as the others.  From what I've seen of them, they're all about planning one project.  The interface I describe will plan only one project, but our system manages 100+ active projects at any one time, so the interface needs to give feedback as to whether you're allocating someone who's already allocated elsewhere.  It's also the way my colleagues have been used to planning projects, so we can't change the interface too radically (to the calendar style, for instance). We have something similar to it already, but it's just a series of input boxes for weeks going forwards (and is consequently terrible to use).  We only just managed to ween them off Excel, but they're seriously pining for draggability. ;-)

And also to clarify, when I said "how much" I meant $/hr.

Thanks again!

-- 
Tim McEwan
Sent with Sparrow

[Clifford Heath]

It's not cheap, but surely if you're already using MS Project, a
distributed MS Project
Server would be cheaper than building it yourself? <http://www.microsoft.com/project
>
Or doesn't it handle cross-project resoursing like you need?

The possibility for scope creep is huge in a system like you propose
(public holidays,
vacation schedules, integration with personal calendar systems, etc).
Been there,
done that (built a GANTT chart widget in the 90s for Telstra to use in
some scheduling
system - luckily I only had to build the widget, not the whole app!)

If you want to build it anyway, the graphical elements of the UI would
work best if
built using Raphael. Raphael is awesome, and doesn't require mad JS
skillz to
make it sing, though I've found they help :). All the drag&drop
capability you need
is either already there, or you can add it at http://github.com/cjheath/Raphaelle/

Not willing to hazard a firm guess at what price you'd get those
skills though. Anything
from $60-$120 an hour, and the rate won't necessarily correlate with
the productivity :).
I wouldn't try to push it down to $40, or you'll probably get a nuff-
nuff.

Clifford Heath, Data Constellation, http://dataconstellation.com

[Nicholas Faiz]

Just a few thoughts on this.

For time estimate, you should spend a lot of time working on your user
stories, and have a firm representation of the system you want to
build in a concise format. Having a legacy system to point to can be
problematic, as it's always ambiguous what to take forward and what
not to replicate. You can ask your guru hired dev to time estimate the
stories, which is much easier.

On top of that, you have to set up servers, practices, etc.. 2 months
might be possible, but it sounds improbable. Better to have a good
description of the basic system you need to accomplish your goals, and
then estimate that and add another half again (in case things blow
out) to be safe.

A guru level coder would start at $90 an hour, at least, I'd think.

HTML 5 and JS sounds great. Also, who's the design expert? If it's not
a coder, then making sure you have your design templates in place
(which can reflect your user stories) can really speed things up.

Cheers,
Nicholas

[Ben Schwarz]

It sounds to me like you've got a solution before you've undergone a design process… 

with that in mind I'd suggest two simple things:

1. take a step back and ask yourself what problems you're aiming to solve and take a stab at how you think you'll achieve them.

2. 3 months is a lot of software… have you prototyped anything? how far along towards your goal can you get in 1 week, 3 weeks? 

Having said all of that, David Goodlad (a Melbourne rubyist) had a prototype of a really interesting scheduling app, perhaps he would be interested in licensing it to you? 

Good luck!

[Jason Crane]

You might also want to engage a User Experience person to validate
that what you think you're building is what your users actually
want/need. Depending on how much time you have them on board for,
their outputs can directly feed into your stories. Then - as you work
through the stories they can design the interface and run user testing
sessions to ensure the build is meeting expectation.

J

[Daniel]

For a reasonably simple 'plumbing' project for <500 users, is a guru
coder the right choice (hard to hire/keep, expensive)?

Also, would writing a good training manual be cheaper than making the
system so usable that users don't need training?

- Daniel

[Tim McEwan]

Hi all,

Thanks for the thoughts and assistance - much appreciated.

@Clifford: we aren't using MS Project and ~35 of our 60+ staff and students are Mac users - so it gets really cumbersome when you're tied to a PC-based solution.  (We're coming from Filemaker + Excel.  Filemaker was chosen b/c it was cross-platform, but then they tied it to Excel via ODBC on PC, which they could never successfully replicate on Mac.)

@Nicholas: you're right, there's still some behaviours that are ambiguous so I need to story it all up.  Also, this is meant to be a replacement component for an in-house project management system that's been live for over a year now, so we have production, staging & CI servers already setup.

Design expert?  Ha!  The push is for functionality right now. :-)

@Ben: that's interesting - I haven't studied SE or even CS (I have a nano degree!) so I'm kind of winging it.  Early on in the user-consultation phase I asked people to try and forget about everything and tell me how they plan, etc.  But we've got an established history and systems that people are familiar with - from here and other places.  When it comes down to it, we have to allocate people to tasks and time to people - while dodging leave, their work on other projects, varying availabilities, etc. - over a large time span (making calendar formats a little too unwieldy), there's only so many ways to do that, right?  In the end it became easier to ask them what they'd change about the current system.

@Jason: being the in-house sysadmin, I get feedback from my colleagues every day - if it's not up to snuff, they will let me know (loudly and frequently).

@Daniel: I'm after a guru coder to work closely with so that I can learn from them.  As I'm a permanent employee of the institute, training me up adds value.

And there's no way a training manual would cut it - it's not that the current system is complicated, it's that it lacks functionality.  We have some staff going to back to Excel to plan and having a casual copy the numbers into the web interface. :-(

Thanks again all!

-- 
Tim McEwan

<snip>

[Mike Bailey]

Sounds like a great project in that:

- the customer is not just on the room, he's hacking on the code

- you're got an existing tool and system in place - much easier than starting from scratch

- it's an internal project - users can be exposed to it before it's polished

- you have an existing userbase - tight feedback loop

Im guessing you'll be able to make decisions on what features do and don't get done.

DHH looked back on Basecamp and said it wasn't actually Rails that made him so 

effective, it was having autonomy in making design decisions. You can get a lot more flow

from a project you're close to. Best of luck with it! :-)

- Mike

--

[Ben Still]

I agree with Clifford- MS project server might be something worth
looking at if "MSP on Rails" is what you're after, and +1 for
Raphaelle

I was an avid MS Project person until a few years ago, and I think for
some reason MSP has become some kind of "default" for how you might
manage a project. I used MSP for pretty much everything, but with the
benefit of hindsight (IMO) it is pretty inflexible. It provided a
false expectation of how long a project might take and how it's going.
As a piece of software, it hasn't really changed much in 10 years (OK
the buttons are shiny and icons are new). We now rely much more on
stories and use Jira and Greenhopper for managing and scheduling
tasks. Tasks in the real world don't neatly line up and change from
"todo" to "done" - they need to get worked on, QA'ed by someone,
perhaps a bit more work, and then closed. They have constraints and
blockers. My point is that MSP or Omniplan might not be the best thing
to model your app on.

Other tools and ways of approaching the problem might better suit your
actual needs. For example we looked at what information was meaningful
for our team on a daily and weekly basis- turns out no one actually
understood the Gantt charts I was making. Now everyone gets an email
each morning with their tasks/whats on for today & this week, plus
some other similar tasks if they run out of stuff to do. We've also
made some stuff to pull information out of Harvest and Jira and show
this on a wallboard http://redant.com.au/blog/the-story-behind-our-wallboard/

Its always fun to build something new from scratch. But in this case
you might get further along by finding an existing tool (there are
thousands out there) and then building something that works with that
if you need to. Or to Ben Schwarz and Jason's points, at least take a
step back and see if MSP is what your users actually need.

regards

Ben

On Jan 31, 6:54 pm, Clifford Heath <clifford.he...@gmail.com> wrote:
> It's not cheap, but surely if you're already using MS Project, a  
> distributed MS Project
> Server would be cheaper than building it yourself? <http://www.microsoft.com/project
>  >
> Or doesn't it handle cross-project resoursing like you need?
>
> The possibility for scope creep is huge in a system like you propose  
> (public holidays,
> vacation schedules, integration with personal calendar systems, etc).  
> Been there,
> done that (built a GANTT chart widget in the 90s for Telstra to use in  
> some scheduling
> system - luckily I only had to build the widget, not the whole app!)
>
> If you want to build it anyway, the graphical elements of the UI would  
> work best if
> built using Raphael. Raphael is awesome, and doesn't require mad JS  
> skillz to
> make it sing, though I've found they help :). All the drag&drop  
> capability you need

> is either already there, or you can add it athttp://github.com/cjheath/Raphaelle/

>
> Not willing to hazard a firm guess at what price you'd get those  
> skills though. Anything
> from $60-$120 an hour, and the rate won't necessarily correlate with  
> the productivity :).
> I wouldn't try to push it down to $40, or you'll probably get a nuff-
> nuff.
>

> Clifford Heath, Data Constellation,http://dataconstellation.com

> >>http://www.queness.com/post/656/10-beautiful-jquery-and-mootool-calen...

> >> For more options, visit this group athttp://groups.google.com/group/rails-oceania?hl=en

[Clifford Heath]

On 02/02/2011, at 9:44 AM, Ben Still wrote:
> I was an avid MS Project person until a few years ago, and I think for
> some reason MSP has become some kind of "default" for how you might
> manage a project. I used MSP for pretty much everything, but with the
> benefit of hindsight (IMO) it is pretty inflexible. It provided a
> false expectation of how long a project might take and how it's going.

Totally agree. I use a deep task/goal nesting, where you can estimate,
work and assign iteration/phasing at any subtask or supertask level.
Many tasks get "done for this iteration" while still having things
outstanding
for a future one. This was especially the case when working on code that
had to be implemented (differently) across many platforms. You need to
handle the fact that software is never "done" and often, neither is a
task.

In the late 90s I extracted the schema from MSP Server's SQL database
and used it as one of the seeds for a new web-hosted design, but that
never got funded. I think it could have been an excellent product
however.

> We now rely much more on
> stories and use Jira and Greenhopper for managing and scheduling
> tasks. Tasks in the real world don't neatly line up and change from
> "todo" to "done" - they need to get worked on, QA'ed by someone,
> perhaps a bit more work, and then closed.

In other words, each task is a mini project by itself. How well does
Jira handle that?

> We've also
> made some stuff to pull information out of Harvest and Jira and show
> this on a wallboard http://redant.com.au/blog/the-story-behind-our-wallboard/

That's really nice.

> you might get further along by finding an existing tool (there are
> thousands out there) and then building something that works with that
> if you need to.

I've not seen one that handles the richly structured hierarchy that I
use and want,
except the indented plain-text files which are my current fall-back.
Truthfully, a
web-based solution would have to be pretty good to be better, for
small projects.

Clifford Heath, Data Constellation.

>>>> oce...@googlegroups.com.

[Tim McEwan]

I see where you guys are coming from, but we're a scientific research institute - not a software dev house. :-)  We don't work in iterations and are only concerned with weekly resolution, not daily.  I'm going to read the wallboard article properly tonight and I'm always on the hunt for tips from the software realm to bring into our project management, but I have to take it slowly with these guys. :-)

I've got my manager & director using Pivotal Tracker fairly readily though.  Win!!  (Thanks @mocra for getting us off Redmine.)

-- 
Tim McEwan

To unsubscribe from this group, send email to rails-oceani...@googlegroups.com.

[Tim McEwan]

Ooh, I almost forgot about SproutCore and Cappuccino.  Can anyone comment on their suitability for such a purpose and perhaps compare to Raphaël?

Thanks!

-- 
Tim McEwan

[Scott Harvey]

I've seen some really slick applications written using Cappuccino but
the major problem with using it in your situation is that it is
written in Objective-J which will make it difficult to find people in
the future to maintain the application unless you have someone on
staff that can do that.

SproutCore might be a bit more suitable since it is just standard
Javascript but when I have spent time trying to write something with
SproutCore it has always seemed a clunky with documentation not really
the best but that may have improved recently.

If you are thinking about building a client side heavy application I
would be leaning towards using Backbone.js or JavascriptMVC to
organised your javascript and then on the backend you can have a Rails
application.

Scott Harvey.

[Ben Still]

re: not being a software dev house
yes- I agree. But my issue with MSProject isn't really software dev or
iteration related - it is more that as a model it creates very
misleading information. Each task is relatively simple- and this tends
to give a false sense of the size of a project. Visualising this as a
Gantt chart (IMO) provides very little information other than a pretty
chart that tells you when the project will end (typically on the day
that the project manager wants it to). A MSP Gantt chart of a project
going seriously bad and another that is fine looks pretty much the
same unless you know what you're looking for. More critically, I found
that no one understood any of the information in the charts I was
sticking up everywhere. It's not that they aren't clever people- they
just didn't care as they had lots of their own stuff to work out.

My point is that I'd avoid trying to follow the MSP model. I would
suggest:
* Think about what kind of information your team needs to be effective
and how to get it to them. It's unsexy, but email has worked for us
* Consider what feedback they need (how are we tracking) and what is
useful/meaningful. eg: we found people were interested in things like
who was the most busy. We stick this on the wallboard
* I'd break your project into three parts - the stuff each person
needs ("what am I doing?"), team feedback ("how are we going?") and
stakeholder report ("how are you guys tracking?").
* Try to avoid putting all effort into an awesome stakeholder report
with all the whistles when a) they probably don't care about how
awesomely interactive it is, and b) if it is non-useful for your team
it will probably whither and die
* Sounds like we had a similar issue with a portfolio of projects
rather than just one. We looked at what the implications of this were
and how to solve problems we were having. eg: we're capable of working
on several projects at once, but do we have the capacity to take on
another project next week?


@Clifford - I'm not sure about Pivotal Tracker, but Jira allows each
issue to have custom fields and triggers, so you can stick as much
information in there as you need. Certainly much more useful than the
MSP schema of "you get 10 custom integers and 10 text fields". For
each project you can have a set of issue types, each issue type can
have a different schema and follow a different workflow if you want. A
workflow can be as simple or as complex as you need. Each step can
have different triggers (if this happens, send an email to everyone in
group A and send this issue to QA), plus there is an API which you can
use to manipulate issues & projects as well as pull information out to
use elsewhere.

re: wallboards - you might get some inspiration for ways to present
team information here
http://ultimatewallboard.com/entries
in particular the Vodafone one http://ultimatewallboard.com/entries#89095
Also, you might be interested in http://www.geckoboard.com/ if you
don't want to make your own

regards

Ben

On Feb 2, 11:52 am, Tim McEwan <t...@mcewan.it> wrote:
> I see where you guys are coming from, but we're a scientific research institute - not a software dev house. :-) We don't work in iterations and are only concerned with weekly resolution, not daily. I'm going to read the wallboard article properly tonight and I'm always on the hunt for tips from the software realm to bring into our project management, but I have to take it slowly with these guys. :-)
>
> I've got my manager & director using Pivotal Tracker fairly readily though. Win!! (Thanks @mocra for getting us off Redmine.)
>
> --

> > > this on a wallboardhttp://redant.com.au/blog/the-story-behind-our-wallboard/

> > To unsubscribe from this group, send email to rails-oceani...@googlegroups.com.