2025-10-15 Yes, all regularly scheduled team mentoring sessions are on for today. |
2025-10-13 The pace quickens as we try to converge on solid product definitions. Broadly speaking, we have not seen anywhere near enough experimentation, piloting or exploration of design alternatives yet to feel confident selecting one way forward. And indeed, that's the goal: confidence. It all gets very expensive to just gamble, find we guessed poorly and then need to either walk it all back or ride in with a substandard product that probably fails in the end. There is a lot of work yet before us. Remember, ideally all 'green light' grades will be somewhere in the A range. There is no merit to approving a spec that we know contains fundamental flaws or gaps, since then a team goes forward to build a flawed product. We need for you to have opportunity to build a good quality product. So again in the ideal case, grades at this point don't vary so much as the number of edits (and time) needed to get approval. Let's get in the game. |
2025-10-10 While working on the green light document, I suggest at least one team member peruse the report-template.docx which is used in final delivery. Not everything mentioned there is required now but my bet is that some of what is required then would be convenient to save now as we will have them right at hand. Design alternatives would be one such thing. Don't make yourself have to scramble to reconstruct such things later. |
2025-10-10 Effort continues to be an instructive perspective on velocity for those who consider it. |
2025-10-10 Now that the dust from yesterday's demos (and subsequent deep dive meetings with teams) has had a chance to settle, let's reflect a bit. A couple of these projects are all of a sudden much more focused. Not entirely so but better. How did this happen? I'd like to make the case (again!) that this is due to modeling. Any time we have freedom to substantively discuss properties of our product we discover important properties that were missed back when we were just talking about them 'out loud' and with vigorous hand waving. Huh. The actionable info there is what? How about that we should proactively figure out ways to stimulate even more modeling activities of substance. Refactor, improve and repeat. Fast. Trying to get to even the current point of understanding but only in code would be a mess, wouldn't it. And then we'd feel pressed to use it rather than walk it back to improve. My homily for the weekend is this: don't push back against modeling techniques, lean into them and improve your products. |
2025-10-09 Today led off with quick team demos, followed by a compact pitch about how one tests for value, not just functionality. The latter is essential for our acceptance test plans on the upcoming green lights. After class a lot of spectacular discussion with several teams on their many design alternatives coming up. Now is really the time to pour it on! Don't just come up with a solution approach - come up with another, and another, each time with evaluation of its costs and benefits. We are supposed to be doing active discovery so go actively discover. |
2025-10-07 Today led off with team reports as the pace on projects picks up. We discused at length the role of modeling in trying different solution approaches on for size, and covered the major components of our class product definitions (our "green light to build.") Let's underscore that the plans should address what support tools and resources we plan to use in the projects. We're not using outside resources with prior arrangement. (Immediately setting up a github project is not your best choice. Presuming that our customers will all be using free student accounts on ChatGPT or the like is also just as sketchy.) If you need hardware purchased (not naming names here...) then waiting to the last minute and expecting things to appear immediately is a situationn awareness fail. Some of these things take time. Plan accordingly. We had a couple of great and spirited exchanges about whether a customer want this or that functionality. This was fun! And it underscored the inadequacy of trying to sort it out with hand waves and imagination. The topic of the day again is modeling, and that is what we would use to articulate one or the other of these solution paths. This lets us explore them in more detail for cheap, and then the stakeholders know in more concrete detail what it is to which they either agree or disagree. |
2025-10-04 Thanks to the majority of you who tried the practice exercise of creating a product definition on TerpAI. We will discuss more on Tuesday, but for now a few thoughts to put it in perspective. First, why are we doing this? Well, remember, most of what we are doing in these initial weeks of the class project involves discovery, analysis and planning. And what is the output of that activity? A good product definition - the expression of how we intend to solve the problem. So it seems like offering an exercise like this in the small would make sense. And remember also, we provided a number of actual specs of previous projects, so you have examples of what you must provide. So far so good? The TerpAI exercise is an opportunity work the kinks out on expectations. What I see in the product drafts is a huge range of qualities. Some are elaborate and cover most of the criteria we might have had in mind, and some are ... thin. A product spec in one page seems pretty unlikely to capture enough information that someone could build from it, right? That unlucky implementation team would have to pretty much do all the basic research for themselves anyway, negating the merits of having tried to work it out in the first place. Kind of a waste. To be clear, those thin specs would not fly for launching the build phase of a project in this class. Here is how to get more substantive feedback for yourself: try the 435 102 bot which is other agent visible to you on TerpAI. This is configured to analyze product definitions as created with help of the 101 bot. We discovered one wrinkle along the way, which is that no matter what it assures us on upload, it doesn't understand a Word format, so convert your terpai-spec.docx file to PDF and upload that. Then just prompt with "Analyze this file." You should get a summary of key points that we'd look for at top level. And you can bet those are criteria we will look for in the class project submissions that will be drafted soon. |
2025-10-03 Effort paints an interesting story for those who care to read it. |
2025-10-03 Now is a great time to reflect as a team on interaction with our visitors yesterday. Do this while it is fresh, and don't be afraid to adjust and improve the solution approaches you may have already had in motion. I bet these will continue to evolve dramatically in the next couple weeks. (Change is what again?) We don't invite critique in order to argue with reviewers and defend original decisions - we revisit decisions, stick with ones that we can still support with objective reasoning and blend in better decisions that the reviews might have inspired. Remember our timeline. We need to have an approved plan in no more than two weeks. In order to have confidence in product decisions, we'll need to see substantive demos or experiments next week, since in the spec we don't care about opinions, we very much do care about informed opinions. Let's inform ourselves with active discovery as has been discussed for several classes. On Tuesday we'll remind some of the high points of the spec process, and on Thursday we might leave some time for each team to give a demo of your pilots, experiments and risk reduction activities. Plan accordingly. Insofar as the spec document itself, remember we offered many examples of such in the class repo, and you can (and should) practice writing one using the TerpAI bots. (Though I notice only five students have done this. Your spirit guide is always sad when he offers tips to which nobody listens.) In practical terms, what do you think is the pace in that last week up to the deadline? Kind of packed maybe? Yes. Expect that the volume of emergency spec review requests will go through the roof; banking on immediate turn around with comments from your spirit guide is not the wise move. |
2025-10-02 Thanks to the veterans of 435 who generously donated their time to advise project teams today. Let me know if you have any follow up questions - I am happy to get you in direct touch if you didn't already grab contact info today. |
2025-10-02 Please sit with your teams today and be ready to rock promptly so we make best use of the time offered by our visitors. |
2025-09-30 Discussion today on requirements expression. We included team reports as the projects are still getting launched, plus a bit of team time. |
2025-09-30 We are assured that the 435 virtual client is available to students on TerpAI this morning. |
2025-09-29 Well it looks like those in charge of pushing our 435 agent to TerpAI hasn't yet gotten it done. Happy if any of you let me know the first you notice it but in mean time I have bugged the powers that be and will plan on updating you here once it is available. |
2025-09-29 The Q7 poll is open now on the mentors site. Please get full team participation. We'll harvest the data early Wed. |
2025-09-26 Nice to see good mentoring tips offered to team members this week, but I notice still many comments which are speaking in third person as if to me. Remember the goal now is not to report to me what people did, it is to coach your team members to help them become better team members. So ... speak to them in these comments. Now that team members have shared, take a look at what they did say to you. Are there good tips for how to up your game? Can you make a better team accordingly? What do you learn from them? Effort this week picked up after our discussion about same in lab on Thursday, though still not full participation. Remember, tech is the easy part; the project is made or broken based on how well we get everyone into the game. |
2025-09-26 This post describes the TerpAI exercise which we brought up in lab on Thursday. Based on this initial project statement we'd like for you to interview our virtual client (at TerpAI) and use that information to draft a proposed product definition along the lines of requirements of our timeline document. It is okay to leave out the cost estimate and press release at this time. The intent is that this be effective practice before going into the drafting of our product definitions on the live projects this semester. Hopefully this puts on your radar the kinds of issues to be alert to (and take seriously) before making impactful decisions on your class project. The virtual agent should appear as one of the bots available to you in your TerpAI dashboard at some point Monday morning. Study the project statement and interview the agent. Save the dialog from your exchange with the agent and push it as a text file "terpai-dialog.txt" to your repository folder. Create your draft product definition as Word document "terpai-spec.docx", also to be pushed to your repository folder. I will pull these for evaluation starting the morning of October 4th. Yes, this is an individual exercise. |
2025-09-26 The so-called Q7 survey as discussed in lab will open on the mentors site (look under Polls tab) early Monday. It is quick to take and we get best value from full participation on teams. This will be harvested early Wednesday. |
2025-09-25 Lab with updates to how each team is bootstrapping their projects. |
2025-09-25 Sorry to see so few goals statements updated to express testable objectives. So far this week the launch velocity overall seems ... low. |
2025-09-24 A couple thoughts as we quickly ramp up on the class projects ... Now is perhaps a good time to revisit some of the tips from previous semesters as shared by students. (I sent that to you as part of the class spam before start of semester, remember?) Students offered tips about how to succeed in projects. Maybe they shared something you can use. Are you trying 'new stuff' as discussed in class? Or reverting to the generic work practices everyone used before starting 435? If you want the generic outcomes then be sure to use those generic practices. My best advice though is to humor me. How about trying stuff we have been teaching? The tip from an old guy who has done this before: There is no more important time to gain ground on a project than at its start. A month from now your future self will feel pretty crunched no matter what (I guarantee it), but time you burn now is time you won't get back. Do your future self a favor. Early steps set our velocity for an entire project since our expectations for interaction rock up pretty quickly. To make all that a little more concrete, let's review: Next Thursday I have many veterans of 435 returning to offer reactions to your solution approaches. We probably ought to have a solution approach to pitch, don't you think? It won't be something carved in stone. Nobody is making a blood oath that this is the solution approach you must use. But there is nothing to make better in the session if there is not something substantive to start with. This in turn probably means you will have had a discussion with the client, had a work session or two to go over options, and of course written stuff down. Working back from there, I hope it is clear what we need to be doing this week. The Terps who volunteer their time away from jobs to which you aspire will give incredible value, but only if you do your part. None will say "here, just do this ..." |
2025-09-23 You know, I saw a lot of teams start to organize after class today. Cool! We need to hit the ground running. But for some reason I'm not seeing contemporaneous notes to the effect in the ticket system. Hmm. How about we get the right habits from day one ...? |
2025-09-23 Discussion today was a retrospective on scrimmage 3 plus elaboration on practices related to the class project. How do we get going? We had first face-to-face opportunities on new teams too. On Thursday we'll illustrate more about we mean by active discovery in a new domain. Remember our ticketing and weekly peer mentoring obligations started now. |
2025-09-21 Semester projects are now assigned. Check out your personal folder for the tasking info, and there you'll find a link for the shared team folder to support these projects. While we use a variety of resources in all these projects, please keep all the graded and PII info limited to this folder. Much of the code we might maintain elsewhere will be shared and extended later, and nobody wants an inappropriate data spill. We have practiced meeting one another in different ways, and now we do it for real and see with what effect. First, please update your goals statement (per next blog entry), but then let's also plan to come up with a team charter by Friday. We will discuss this in class Tuesday. This is not huge, but it is important and at a minimum will include practical information like times each week the team will have an all hands face to face, how to do daily touch base and our expectations for performance. To be abundantly clear on the assignment, our timeline document is now in effect, as are our expectations for logging tickets and weekly peer mentoring. To say the timeline simply: You have four weeks to plan what we will build to solve the problem (and get me to agree to the plan.) You will have four weeks to build the product, and then convince me it is a complete build ready to put in the hands of our customers. And then you have almost four weeks to observe your product in customers hands and, once you are sufficiently horrified at the results, refactor the product to what you should have done in the first place. Deliver that. |
2025-09-21 At your early convenient opportunity this week, please update your goals document to ensure your objectives in this course remain clear and also (now) are stated testably. How will you recognize you achieved them? Previously you updated to also state how your bring your superhero powers (Gallup Strengths) to the game in order to advantage your team, so this builds on what was written before. And yes, this is me being clever: I would love for you to consider your strengths and goals at the same time you are establishing community with your new project team for the semester. Yes, all these pieces dovetail together! |
2025-09-18 Chiefly lab today with disscussion on ticketing and peer mentoring practices, then plenty of team time in prep for scrimmage 3 delivery tomorrow. |
2025-09-16 Discussion today on principles of software engineering that we will practice applying on the projects through rest of the semester. |
2025-09-15 Welcome to the new work week. First reminder, glad to see the 400-level course surveys come in but we are only about half way there. This is something very useful to tasking class projects, so please don't make me fly blind. The sooner we get full participation, the better, thanks. Next, this week we will invite you to do a review of team performance on the scrimmage project. These sorts of reviews are obligations you will conduct throughout your career. We practice the mentoring throughout our class project (more on this later) but this week we will practice with a review. Specifically, I've set up our mentors site with the current groups and invite you to offer me an assessment of each of your team members - including yourself - by logging in (usual campus directory ID) and going to the Advice -> Add tab. Entry is live - no separate "submit" is needed and you can change or update it as you like up to the deadline. Ideally this is something you'd do after the team submits the project on Friday, or at least after the bulk of the work is done, so we have the best input, but in any case I will harvest the data at 6PM on Friday which is after scrimmage 3 is due. In the free text box, offer a simple description of what the team member did and how well they did it. Please be clear and specific. Also, please share your thoughts on working with each team member again. These data are only for me; they are not shared. Later we will have a weekly obligation to offer mentoring tips to team members, and those data will be anonymized and shared. But that starts next week. Historically these reviews have been very well received by students, since at this point in our careers we often do not know much about how we are perceived by others. This is a way to get tips on how to up our game and make an even more effective team. The site is open now in case you want to play with it, and you can (and should) update it with the final values on Friday. |
2025-09-13 Scores for scrimmages 1 and 2 are posted. See the grading summary in group folders for more detail. The mission now for scrimmage 3 is to follow the plan you just laid out and get us the winning product. Do be specific about reflecting on how you planned as a team and track how it works out. This is the only way we become more proficient in such things. The grade server is up to date and 21 course points are accounted for. If you feel nervous about the last day of schedule adjustment coming up (Monday) then you can compute for yourself the likely trajectory from here. |
2025-09-11 New survey released! This morning I pushed a CSV file to your personal folder, showing 400-level courses taught in CS in recent years. Please go through this put an X in the first column of all such courses you have taken successfully or are taking now. Save it and commit it back in place by SOB Monday morning. These data are invaluable when it comes to crafting effective teams to pair with the class projects coming up. Why not just take care of this today? |
2025-09-11 We explored process models, processes, planning and related practices, with heavy emphasis on why such things matter. Prototyping came up along the way too. Then over to some lab team time. |
2025-09-11 A PSA for those keeping tabs on such things: The grade server remains up to date, and 11 course points (in the race to 100) have been accounted for to date. My goal is always to ensure there is graded material returned to students in advance of the campus schedule adjustment deadline - Monday in this case. By early Friday we will have accounted for 16 points, and I have every hope that I'll get scrimmage 1 graded by mid day. Scrimmage 2 is due late in the day, putting us 21 points in. |
2025-09-10 How about those updated IDP's into the repo? |
2025-09-09 Class today discussion about what is in our scope of study in software engineering, which we approached by going over ways we know it has not been done well in the past. We started to motivate consideration of principles of SE, which we will continue with moving forward. We gave the balance of time over to discussion about what to put in your plans for the scrimmage. What goes in your plan? Once you have had a chance to think about it (and submit as scrimmage 2) I'm looking forward to comparing plans to show the many different things people might put (or not) in such a document. Later of course we will see how those plans worked out, by means of scrimmage 3. |
2025-09-09 Happy birthday, computer bug. |
2025-09-07 At this moment the grade server is up to date, 9 course points (basically one letter grade worth) are accounted for, eight scrimmage teams are assigned and we have a weekend to reflect before starting fresh in the new week. To assist in reflection please consider these supplemental materials. Quality improvement is an on-going game after all.
|
2025-09-06 A reminder going into the new work week that if you haven't yet submitted materials from the first individual exercises (goals, Gallup, resume) then you should get them in so right away. We will continue to rely on these as we move ahead. |
2025-09-04 A lab today with plenty of discussion about what makes a good team member. We gave some things to try, we discussed at length what we bring to the game (aided by Gallup materials) and we are off and running on the first scrimmages due next week. Remember to followup the practice today with a deeper dive into your Gallup Strengths, and then upload the expanded IDP to your personal folder in the repo. |
2025-09-03 Glad to see activity in the repo as you get on top of the first assignments, and similarly glad to see teams forming. If your schedule allows then consider coming a bit early to class on Thursday (or stay after) to schmooze with other members of the class. Let's build those connections. As you identify your textbook and support materials, plan on diving into the initial 'big picture' chapters. All the suggested materials will have the same organization at this point so wrap your brains around the overall mission in SE and be sure to have a book-learning sense of what is a 'process model' and why we care. |
2025-09-02 If you had trouble checking out the class folder in our repo then try again. I had missed a line in our configuration file. Hopefully fixed now. |
2025-09-02 Plenty of action items coming up. Let's get these in the queue early to help you plan effectively. First, an individual exercise to practice simple tasks in the small (for cheap) in order have confidence we're heading in the right direction later when things become large (and expensive.) Mailtools is a Linux-based system created in a prior class. 435 students should be able to access this project in our repository at https://vis.cs.umd.edu/svn/projects/mailtools/src (using appropriate SVN tools, of course.) The simple (individual) exercise is to tell us about its "Space Invader" feature. Who was its author, and how many commits are related to it (just this feature, not the full Mailtools product)? Don't agonize over answers; derive reasonable responses for each of the above questions, place your answers in a file "svnfun.txt" and check it into the root folder of your personal repository by 0700 on Wed, September 10. Scrimmage 1 (the first group exercise) is a two-foot putt: By 0700 on Friday, Friday, September 12, set up your assigned VM as a web server. Put up something that would convince a skeptical visitor landing on that site that the server relies on a database. Ensure your VM itself evidences basic care, maintenance and security. (Some students are cyber mavens - cool, so to be clear, this is not an obligation for you to go nuts with defense!) Per convention for group assignments, please submit a cover sheet, according to the conventions for digital signatures with the template in your group folder. This exercise serves as a forcing function to get everyone on a group in some form or other; it lets us practice conventions for projects; and it is a good opportunity to practice following our "do the right thing" directive. Heads up! Scrimmage 2 will be due at COB that same Friday. This is also an easy target (problem statement is below), but will require talking as a group in order to sort it out. Write a "plan" for how you intend to solve the problem. The deliverable is your plan in Word document as "scrimmage02.docx". It should persuade me as an ostensible product manager that you are on track and can solve my problem. (Cover sheet: yes.) Scrimmage 3 is due at COB the following Friday (September 12). For this just follow your plan. Foreshadowing: by that point we will have surveyed all sorts of simple ways that a product can fail for want of technologists anticipating what users actually need. Your mission is not to replicate such defects in this exercise. Don't just write a program; solve the problem. (Cover sheet: yes.) The problem - a capability we'd like to have: We'd like to have an easy way to visualize (read: "graph") data from a collection of uniformly-structured spreadsheets. There, how's that for simple? Give me what I need in order to succeed. Okay, a few more details are probably in order... The starting point for this tool is a set of "uniformly-structured spreadsheets" - what are those?
And what do we want to do? Enable the user to select and graph one or some of the properties over the whole set of input files. If sheets capture weather then we might want to identify a TEMP row name and value field(s), then our product will select those values from the set and pop up an appropriate display. If it was stock values, then we'd identify one or some stock ticker values (if they appeared as the labels in each sheet) and display the graph of those over time. If sheets were about student performance then we'd off the list of student names (again presuming these appear as labels in each sheet) and see the trend of values over time. Some considerations to put on your radar early: First, we will be pretty concerned with usability and work flow. A system that technically allows visualization but with high-overhead data entry costs is probably a non-starter for us. Next, we are interested in rich display options; flexibility in what we name and how we select them is important. And as noted, we want something more than a toy; let's do this at scale. Finally:
|
2025-09-02 First day of class. We spent discussion today setting expectations of one another for the semester. Our action items for Thursday are as announced earlier. So keep interacting with one another to form a team; send me the directory IDs of your four person team so I can issue you a group repo which also gets you access to your VM. Feel free to get settled in there in anticipation of our first simple group project in the coming week. Some random tidbits from today's class:
|
2025-09-02 And we're under way! Repository credentials will have been emailed to your address of record as of Tuesday morning, so per usual convention, if this is a surprise to you then check your spam folder or contact me soon so you can get set up. The initial assignments are due as detailed in the 2025-07-11 post. We'll see you later today in class. |
2025-07-11 A new semester is just around the corner. We're sharing the the first assignments for those who want to get an early start.
|