Table of Contents
Follow up Questions
* **You mentioned that you're available to meet with your capstone team frequently - would you be
able to (or want to) meet with the team once a week, once every two weeks, etc?**
* **Along those lines, do you prefer to meet in person or remotely?**
My preference is to meet in person with the team weekly. I find that keeps the project moving along well.
I am flexible and open to discussion if the team feels something else works better.
* **Would you be available for quick questions via email, IRC, etc if any big design questions came
up mid-sprint?**
I am available on IRC or other for mid-week questions. I prefer IRC over email as my email gets bogged down.
* With what frequency would you like to see progress demos?
For progress demos, I would like to see us create a project schedule and we can put
demos/goals in that schedule. I don’t feel the need for a detailed project schedule,
but I do feel a high-level schedule helps keep the team on track.
* Would you have a preference on where / what hardware the web app would run during
development and later for live demos?
For the live demo, we most likely need to run on a server at MIT/LincolnLabs, but for development
purposes, we could use a server I have access to - likely a machine named barista.cs.pdx.edu.
* For the Google watch concept, I believe you already mentioned this, but would you be
providing the team with a watch for development if we evaluated the idea and decided to go for it?
Yes, we have funds to purchase a couple watches for the team.
* **You specified that the goal of the web app is to emulate a real-time ICU monitoring service.
Would you also see any value in also having an analytic component for past historical data
(graphing trends of data that's streamed through in the past and such)?**
* **My feeling is that the real-time aspect is what will really demonstrate the
features of S-Store, and so running analysis on past data might not be a huge value
add for the purpose of the project. I figured I'd ask, though!**
Very good question. We are actually very interested in combining the streaming data
with historical data. As you allude to, we agree that pure analytics on the historical data
is not the best use case to demonstrate S-Store. However, we do believe a combination of historical
data, possibly with an analytical component is an excellent use case **for S-Store.
- The problem:
- Lots of employees driving lots of cars
- Get them to carpool and build relationships
- goal: 500 employees: 250 cars 1 day a week
- Connecting riders
- Build relationships between coworkers while commuting
- As an employee I can:
- Create a profile with preferences
- Pick a location to meet, days of week to carpool, time to arrive / leave work
- How many people I can drive
- A fun avatar picture
- Create a carpool
- Name the carpool, set the meetup area, set carpool capacity
- Set a carpool profile (talkers, coffee drinkers, sports talk, smokers)
- Driving schedule (rotation, days of week, time of day)
- Statistics (leaderboards, competition)
- Find employees near me who want to carpool (map view)
- Be notified of new carpools nearby
- Request to join a carpool
- Send messages to carpoolers
- Reminders of upcoming carpools
- Opt out of carpool temporarily or permanently
- AS a NAVEX Global team member I can:
- See leaderboards: top carpools by:
- Most people in car
- Most days of week
-
of miles saved per month
- Longest established carpool
- Participation by office location
- See map of all carpools by office location
- App features
- Available anytime, anywhere (web app)
- Fun user experience
- Easy to use on mobile
- Only NAVEX employees can join
- Secure: protect employee info
- Source on GitHub (MIT, GPL)
- Unit tests, functional (acceptance) tests
- Tech stack: open ended: rails, django might be good.
- Web primarily, maybe native Android app.
- Responsive web design
- Must be stable
- Backend stack of choosing (hosting..)
- Company is flexible in how they will host themselves.
- Sponsor schedule: very available, in person and remoe.
- Have stakeholders in HR.
- Branding: allow client to style with their color / logo.
Follow up Questions
Carly – great questions. Here are my suggestions:
Security: A lot of this will be dependent on the authentication methods that
you design and implement. We certainly have some Engineers here that could
review the design and provide guidance in this area.
Customizing for other companies: Our primary goal is to have this working for our
company. The ability to work for other companies is an extremely low priority.
In fact it would be one of the first things cut if the scope of the project was too large.
Meetings / Feedback: I would suggest that we meet at least once a week for a progress
review – we’re flexible on the location, either PSU or our office – maybe we could rotate.
A Skype video call is another option. Beyond that, we want to be highly available to you
for any questions you would like a quick response too. Email is probably best for that.
Doug
- Basic concept
- Open source! Doesn’t necessarily have to be internal
- Multi-platform
- They are flexible on their workflow: they’ll adapt to what we build
- We’d have access to the team to ask questions
- In the beginning: pretty manual process
- Open to actually scanning barcodes, QR codes, etc
- Mobile app: they will ask around office to figure out which platform
- We will be provided with servers
- Software they’ve looked at
- Must run on Windows, bad
- Can’t interact other than exporting / importing CSV
- Can’t access over the web
- Wants hosted solution - web app
- Inventory needs for small business
- Managing physical inventory
- Accuratlly represent data particular to phyiscal inventory of an org
- Provide simple means to interface both at admin and transactional level
- Convert data to flat files for transfer/distribution
- Problems:
- No centralized management
- Inventory not managed. Some items recorded in various places.. need single source of truth for inventory
- No standard operating procedure
- Need standards for handling inventory and logging it
- Need to provide good reasons to take action
- Too many steps
- Current processes are cumbersome. Make the thing lighweight, easy
- These leads to errors, inefficiencies, stress
- Software solution:
- Current software is too expensive, inflexible
- Must be flexible, simple
- Requirements
- Users: access control
- Produces: unique id, name, description
- Batch details: id, item id, produce id, date, job num, quantities
- Product status: location, current quantity
- Product transactions: users, dates, job numbs
- Transactional history (important)
- Historical transaction data
- Application interface:
- User adminL create users / manage perms
- Product admin:
- For users with write privs, add new products
- For users with read only, view only
- Batch creation: add x of y batches with z unique ids for batch
- Transactions: basic add, subtract from inventory
- Optional: mobility
- Mobile app for tracking stuff (add/subtract on smartphone)
- Scanning system
- Discuss scope
Decision Fatigue
Making decisions costs
Can’t get an Android app to capture “stuff” in an inbox
Go back and see what the thing was
Is it actionable?
Decision tree: no or yes
No: eliminate, incubate, or file away
Yes: create projects, multi-step or single step item
What’s the next action?
Less than 2 minutes: do it now
Otherwise delegate, take part yourself
Some do ASAP, some schedule later
Filters
Limiting criteria
What can I do?
What can I do in the time I have?
What do I have the energy to do?
Don’t talk about priority
Context is the real determiner of what you can do
Downtown? Do my shopping
Home? Do my homework
What’s my energy level?
Current best product: OmniFocus for iOS
Android only has Evernote
Limitations:
File centric rather than user centric
No support for the process itself
User needs discipline
Unnecessary navigation
Cognitive load on users
System breaks down when user relaxes discipline
Action filters don’t work
Singly rooted classification destroys context view
No integration of filters: user must construct “engage” view
Trust and reliability issues with web storage
No use of metadata
What I want to do, when I came up with it, when I did something
Collect this data and use it for coaching
Proposal: three buttons: Capture, Organize, Engage
Very visually based solution: see slides
Mechanics:
Free Google Play store app
Windows companion
APIS:
Google drive, tasks, keep
Evernote - great capture, published API, unusable file structure
Can’t record phone calls in Evernote
Extensions:
Location aware, collaboration support (delegation), automatic calendar connections, AI coaching (detect krufty tasks)
Everybody involved in free thing is partner in second one
Open source
Work out licensing for post open source
Real Estate Design App by Ursina Teuscher + Dan
Using knowledge and tools compressed into a simple webapp facilitated by real estate pro
Licensing: real estate has access to thing, trained to use
Both parties fill in relevant info about homes they’re considering
Several parties can work on a decision
Equivalent functionality as a decision table
Basic structure of options, criteria
Slideshow has tabs
Realtor places facts
Purchase takes own notes, uses subjective rating
Combine perceptions in aggregate
Everyone distributes 100 important points
Back end does math, figures out what houses are right
Matrix shape: options x criteria
Display this simply and elegantly
Don’t have to see it all at once
First rate criteria, then rate it
Like Turbotax style?
A progression: later you see facts
Later you see partner’s scores
Facilitate discussions
Useful for Realtors: adds values for their clients, reduce work for realtors themselves
Help keep track of things important to client
Realtor tab shows aggregate of client opinions
Maybe let Realtor demonstrate highlights of houses
Some things can be objectively presented (distances, facts)
Maybe used to rate Realtor (how often are rankings high)
Eventually, realtors use this as a service for clients
This matrix is generated every time clients come to a realtor
Custom built every time
Table begins empty, Realtor adds properties they think the clients would like
Can allow weighting for parties involved (you get double my weight)
Clients have individual accounts
Ideally, this should be executed as real estate only first, rather than general.
Licensing for the interface: IP: not much to be patented. Lots to sell here.
Very flexible in monetary compensation / funding
Estimate our teams effort, develop market rate, we could get money retroactively
Tech:
Web app you can login to everywhere (responsive)
Desktop first
We can figure out how best to translate needs - up to us
No strict rules on tech choices
Hosting modeling:
No plans yet. We’ll think about it.
No huge data traffic right away, but we’ll need to save user prefs and data
Software as a service - cloud hosted