Archive for June, 2011

Cat Juggling

Saturday, June 25th, 2011

I have several projects starting up in the next few weeks, so the challenge of managing several concurrent projects has been on my mind. Managing multiple projects at once feels a little bit like juggling cats:

Even though it seems more difficult, I don’t think it’s inherently any different from managing a single project. Except, it’s absolutely crucial that each project is managed as efficiently as possible.

Being an engineer at heart, my tendency is to micro-manage things. I hate the idea of not having control and not knowing the status of every task all the time. Part of that is my nature, and part of it was learnt early in my management career.

When I first became a project manager, I was reticent to watch over the work of others. I figured, “I asked him to do this task, so it will get done the way I want it done.” The problem was that people did not always follow through and not always the way that I would have wanted them to. It was at this point that a colleague I trusted gave me this short but valuable insight. “Other people are not like me.”

Pretty obvious, but I think we all tend to assume that others will do things the way we would. Sometimes they do it differently and better, and that’s great. But other times, not so good. There were a few projects that blew up because I was not following up how tasks were getting done and so I learnt my lesson.

That’s when I got religion. Instead of assuming that tasks were getting done, I was all over people to make sure if and how they completed the work. Which made me much more effective as a project manager, but kind of a pain in the ass to work for.

I’m now somewhere in the middle, of course, but that does not mean I’ve just split the difference between absentee manager and micro manager. Instead, I have developed my own philosophy of project management that seems to work for me. Here are the 3 tenets:

Get it right early - The first few weeks of a project are the most critical, for obvious reasons. I try to make sure every person on the team knows his role, what are his reponsibilities, how they will work with others, how we communicate, etc. If I can get this part right, then things usually go smoothly from there. The first few weeks are also the time when the project goals, scope, budget go through the most change. Managing these changes and the associated risks, costs, schedule impacts, and expectations are critical to making sure the project is achievable. This is obviously a lot easier to say than do, so I like to engage everybody on the team to help us accomplish this.

Regular communication - This was one area I did not do well the first time around. I’d have sporadic 1-on-1 meetings or I’d only call people when there was an issues. Now, I make sure I touch bases with everyone on the team, and with the customer, at least once a week. It’s not always very long, but there’s that human relationship that needs to be developed that takes repeated and frequent contact, ideally face-to-face. With global teams this is not practical, so we talk on the phone, something that has been made easy and cheap by Skype. At least there is some human communication over the phone. Email and chat is used to keep in touch with quick things, but it tends to be less personal. I’ve found that if I stay in touch, people will bring me issues sooner and I can deal with them before they threaten to sink the project.

The customer, the customer, the customer - Especially within larger organizations that have been around awhile, there develops unspoken organizational and personal agendas and objectives that run counter to project success. Sure, it’s nice not to have to travel if you have a family (or to get to travel if you are young and single). It’s great to be able to work with the latest new tools or on the most challenging 28nm project or work with your best friend. In the end, however, as tough as it may sound, we’re all out of a job if we don’t take care of the customer. Whenever there is a tough decision to be made, it’s important to make project and customer success the #1 goal. That does not mean that everybody has to cancel his vacations and work 80 hour weeks, but it means that the impacts to the customer need to be considered and mitigated. Or else we’ll have a lot longer vacation than we bargained for.

There’s a lot more to project management than these 3 tenets, some of it nuts-and-bolts (how to develop a schedule, do risk management, etc), and some of it very difficult (how to deliver bad news to the customer). I like to think that these 3 tenets provide some guidelines that make the other stuff easier.

harry the ASIC guy

Dunbar’s Number and #48DAC

Tuesday, June 14th, 2011

DAC Badges

My apologies for the recent hiatus in my blog posting. It’s been a difficult time personally for me the past few months, dealing with family illnesses. Hopefully, I can get it going again.

With all that I had going on, it was a relief to escape last week for a few days to DAC in San Diego. After several years attending as a blogger (what DAC calls “independent media”), it was exciting to be on the floor representing Xuropa at the Synopsys Cloud Partners Booth. I still got to see several friends like JL Gray, who wrote up what he heard from us, and Peggy Aycinena, who accused me of being a sellout since I was in the Synopsys Cloud booth and had a Synopsys badge lanyard. And of course, what DAC would be complete without Eric Thune of AtopTech telling me that cloud will never work for EDA. 

One of the downsides of being in the booth was not being able to attend a lot of the other sessions. I missed The Woz, and the Logan & McLellan show, and Gary Smith, and a lot of the panel discussions. I was, however, able to sneak away for the EDA Cloud Computing Panel discussion, featuring the usual suspects and a few new ones. A highlight was when John Bruggeman of Cadence offered to buy John Chilton of Synopsys a beer at the Denali Party and work out a joint Synopsys/Cadence solution on the cloud. No word yet how that turned out. Another highlight was the audience poll at the end where 1/3 of the audience felt that most of EDA would be on the cloud in 3 years. I don’t know if this is correct or not, but this is the 3rd year we had a cloud panel at DAC, and each year the expectations increase. Richard Goering has a good writeup on the panel.

One booth I did visit and get an interesting demo was Duolog. Duolog is a Xuropa customer (you can try out their tool here), which is why I knew a little about them going in. They have a tool called Socrates Bitwise that does register management for processor based designs. In this tool, you specify all the processor accessible registers, their type (RO, RW, etc), the locations (base and offset), and the tool automatically generates the RTL, verification code (OVM, UVM, etc), register package, C APIs and documentation. If something needs to change, you change it in one place in the tool and all the subsequent files are regernerated correct by construction. With many designs having hundreds or thousands of registers to manage, this is a growing problem to be solved. Duolog has a few competitors as well, but their biggest competition is in-house home-grown scripts.

Of course, there were my 150 closest friends I know from years gone by, too numerous to mention, lest I leave someone out. I’m reminded of Sean Murphy’s perfect description of DAC:

The emotional ambience at DAC is what you get when you pour the excitement of a high school science fair, the sense of the recurring wheel of life from the movie Groundhog Day, and the auld lang syne of a high school re-union, and hit frappe.”

An overall impression I, and many others, had was that the show floor was smaller and there were fewer attendees than in the past. The official preliminary numbers, however indicate that DAC was larger than last year, so I’m not sure whether to believe my eyes or the numbers.

For me personally, it was my annual chance to connect with the entire industry, so I got a lot out of it. At a minimum, it provided me with a lot of good ideas that I can work on for the next year.

harry the ASIC guy