菜单

How we build the Tech Radar

01 November, 2018 | 18 min 23 sec
Podcast Host Rebecca Parsons and Neal Ford | Podcast Guest Marco Valtas and Camilla Crispim
Listen on these platforms

Brief Summary

The ThoughtWorks Technology Radar is the must-read report for anyone interested in state-of-the-art software development. In this episode, we take a behind-the-scenes look at how the Radar is built: how technologies get onto the Radar; the in-depth discussions over where they’re placed; and why some things miss out.

Podcast Transcript


Rebecca Parsons:

Hello everybody. Welcome to the ThoughtWorks podcast. My name is Rebecca Parsons. I'm one of your co-hosts, and I'm joined by-


Neal Ford:

Neal Ford, one of your other regular co-hosts. So you've seen the completed radar, but you may not have thought much about how the thing fits together. There's actually a tremendous amount of logistical work behind the scenes. And one of the key roles that makes that happen is Rebecca's technical assistant and the one who's just wrapping up her two year tender is Camilla Crispim, and she's here today to talk about how we put the radar together.


Camilla Crispim:

Hello everyone.


Neal Ford:

And we also have one of the regular tab members who's one of the people, one of the cats that Camilla has to herd, to get things done this morning.


Marco:

Yeah. Marco. Represent Brazil on the group.


Neal Ford:

Well let's go back to the very beginning. So how does this thing to come together? So a lot of people know that all of us gather from around the globe into a meeting space. So how does this come together?


Camilla Crispim:

Well, I think the first step is actually getting people from the group to collect the blips, and they have very passionate conversations with all the ThoughtWorkers that they work with or is in their region. They collect the blips, they do some sort of self censoring to see what they're going to bring to the room, what they're not going to bring. Sometimes they just want to see if others are doing the same stuff that they are doing, but that's the starting point. And then when we get here we put all the post its in the wall and then it's time to get started actually discussing and seeing what's going to be in the radar and what's not going to be in the radar.


Neal Ford:

So if you follow us on social media, you may have seen some of the pictures of all the post it notes on the walls. Those are what we refer to as blips, and there is a fair amount of self-censoring I saw yesterday and broke my heart that there's a Mac, someone has suggested e-macs for the radar, and she had self censored it before she put it up there because she thought it didn't have a good fighting chance given what the competition was. So, so yeah. So how many blips do we normally start with as we go the beginning of this process?


Camilla Crispim:

Usually around 200, that's how we start. And then we still have the existing ones, which is the current edition of the radar, which is about a hundred ish.


Neal Ford:

Okay. So we have all the blips up on the wall. What happens next?


Rebecca Parsons:

And then the games begin.


Camilla Crispim:

Yes. We basically go quadrant by quadrant. So we have four quadrants, techniques, tools, language and frameworks and platforms. And then we start with one quadrant and we go ring by ring, usually from adopt to hold, discussing every blip and the person who suggested that blip starts just describing what it is and why it should be on the radar. Kind of like advocating for it. And people do the voting. And we have an amazing way to get the votes and to facilitate the room.


Rebecca Parsons:

And that's my job, refereeing the technical discussion amongst 20 some odd ThoughtWorkers who are very passionate about technology can sometimes be a challenge. We do have a voting system, we use color cards. Green is yes, red is no and yellow means, "I have a comment." And because there are so many blips, we try to manage the length of the conversation. So this time, for example, I set a width limit on the number of people who could be lined up wanting to say something. And I do very carefully keep a list of who wants to talk because otherwise someone who is more quiet might not ever get a word in edgewise. And there are just so many people, it would devolve into chaos if we had too many conversations going on.


Rebecca Parsons:

So it is very tightly facilitated and we'll try to vote on it. I try to get an early vote because sometimes we are very passionately in agreement, and with 200 blips to get through, we can't be passionately in agreement about everything.


Camilla Crispim:

Yeah, and I particularly thank Rebecca for that kind of facilitation. Otherwise I wouldn't be able to track the decisions and we would have no radar by day after the four day meetings.


Neal Ford:

So would you describe the role of walking the wall?


Camilla Crispim:

Yeah, so I'm actually doing this time for the first time, but it's basically calling out the name of the radar who proposed it and you go after the voting, you say, okay, "It's in the radar, it's not in the radar. It's even what we call too complex to blip.", with when it's a very nuanced conversation or advice and it deserves its own article or podcast or a discussion around it. It wouldn't make to the radar because of the one paragraphs sort of rule that we have. So it's not that it's not important it just radar is not probably the best place for it.


Neal Ford:

So there's a person standing there that pulls the blips off and announces them but that's how it actually gets translated into some sort of format that will eventually become a radar.


Camilla Crispim:

Yes.


Neal Ford:

So someone's taking notes, right?


Camilla Crispim:

Yes, yes. This time Nee, who's going to be my successor is actually taking the notes, tracking the name, who proposed it. And then what's the decision? Is it new? It's not going to be there. So on, so forth.


Neal Ford:

But I think it's important to point out that you guys are not the ones who actually write the content. All you're doing is tracking who suggested a particular blip and that they're going to be in the, writing the content for that.


Camilla Crispim:

Yes, that's a step forward. End the meeting is actually time to start working on the blips, and writing the blips, and pushing people to write the blips.


Neal Ford:

Well that's got to be the easiest part of your job, right?


Camilla Crispim:

Yeah, of course. And we have such a tight deadline as well because we want to get all the blips ready so it can be translated and you know the radar can be on our friends hands as soon as possible.


Neal Ford:

So when you began this role, how long did it take from the time we left the face to face meeting and had collected all this stuff until it was in the hands of these eager technologists who wanted to read it?


Camilla Crispim:

I think it was actually the first time that we had an eight weeks release time for the radar. I think before that it used to be longer than that, and now we are with six weeks.


Neal Ford:

This is part of Camilla's legacy is she had shortened the feedback loop on the radar, because as Eric commonly points out, the longer it takes to get the thing out, the less relevant it is, because we're trying to comment on new technology, so getting it out as quickly as possible is critical, but you have all these people who gathered together to write all these, the content of these blips, these paragraphs. There's time. What could possibly take six, even six weeks to get that? You herd them like cats and get them to submit all this stuff and then it's pasting it into a web page and we're done. Right?


Camilla Crispim:

Yeah. I wish it was. If it was like you just said, it could be three weeks, half of the time, but there's a lot going and a lot of effort going to make the radar go live. We have a whole team, which is highly distributed team, across the globe, to make it happen. So the people from the tab who is in the room, they write the blips. Well actually as they go writing the blips, we trigger what I call like a pipeline, trying to be not to post technical, but it's basically a feedback loop. So it goes through an internal process of feedback, it goes through proofreading process as well, when everything's kind of like get it quite ready, and then it goes to translations. And currently we translate to three languages, Spanish, Chinese and Portuguese. But it has been the time that we actually translate it to kind of like more than five languages. So we make the radar broadly available even if you don't speak or read English.


Neal Ford:

S0 what's so tricky about the translation? Can't you just hire somebody to do that and then be done with it?


Camilla Crispim:

Marco is your go to person now.


Neal Ford:

Exactly.


Marco:

Well, we tried that. In the end we figured out that it's better for a Thought Worker to translate the blips. I guess the way that we write things and how sometimes technical terms, especially in the regional area, it can be translated or sometimes not, so depends on if the industry, it's adopt the term already. Like pipeline. So in Portuguese, we use the term pipelines and usually translation agencies don't know that that what happened in the industry. So we have usually a Thought Worker that does the whole translation, and then we need to review the translations, make sure that whatever was written is the same, I would say with the same passion that was in the regional. Basically we need to make sure that the terms reflect how passionate we are about the blip.

Marco:

And then you have a second round of reviews and then a head of tech does the last check. And then all the translation go to someone from marketing of our side to make the PDF. And now that we have kind of a [inaudible 00:09:21] style of translation, the thing it happens which [inaudible 00:09:26]. So we've got a group, usually the ThoughtWorkers that do the review are basically the same group because you know people are in the project, so they do in the spare time.

Camilla Crispim:

So the PDF is back and then we need to do the QA of the PDF, making sure you know the layout looks good, all the links are in place. So it's a very kind of manual process still. I'm sure in the future's not going to be a very manual process anymore because we don't like doing manual stuff. So as much as we can automate, it's better. And then it's the big day and then we launch it.

Neal Ford:

So let's talk a little bit, go back a little bit about the review process, before translation because it's remarkably consistent given that a bunch of different people that many of which English is their second or third language, put together, because it all starts out in English, and then it gets translated to the other languages that we support. And so once the creators of the blip content have given that to you, what kind of review process does it go through from a technical review standpoint and also like grammar and consistency standpoint?

Camilla Crispim:

Yeah, so I'd say the first big thing is actually just coming back a little bit of what is that exactly what we put out there in the radar. It's basically reflecting our experience. So we send the write-ups as soon as they get ready for our internal community to have a look, and they actually say, "Okay, you could add more detail here or this doesn't sound about right. Could you rephrase that? Because this is what we've been experiencing." So this is the kind of like big thing that I'd say it happens, and then I do some review on the content. Sometimes if its a very controversial blip, Rebecca does the review and then we have a technical writer at Thought Works who does the review as well. He looks after the tone of the conversation, the description, and then we have someone external to actually review all the English sort of content, because sometimes you just look at that stuff so many times that you don't spot typos and anything like that. So we have someone external to do that, who is quite passionate about doing that.

Neal Ford:

So there's an interesting kind of feedback cycle here because the place the blips come from to begin with, our ThoughtWorkers on ThoughtWorks projects, have nominated these things and then it goes through this giant grinder process of radar blip creation and all that. But then it gets back in front of them.to say kind of did we summarize this correctly as to what you're saying before we broadcast this out in the world.

Rebecca Parsons:

The radar is a point in time expression of things that ThoughtWorkers are passionate about, and it's interesting the success of a blip actually getting through the initial presentation and then getting through the cull when we have far too many blips to put on the radar, and so we have to go and take some out. The passion with which an individual advocates for a particular blip really matters, particularly in that cull phase. And so we make no claims that this is comprehensive, but this is an expression of what the ThoughtWorks collective, filtered through this group called the tab, what we're passionate about, what we're seeing, what we see as successful, or in the case of the hold ring, things that we're continually seeing our clients do that we think is a problem.

Neal Ford:

So what would you say the rough number, and we have some rough numbers on this radar because we're right in the middle, directly right in the middle of this radar process right now. So you said we start with about a hundred existing blips and then we add about 200 during the initial phase of picking stuff. But our goal is to get back down to about a hundred, right?

Camilla Crispim:

yes.

Neal Ford:

So that's the culling phase that Rebecca's talking about, that we're about to start going through.

Camilla Crispim:

Yes. At the moment we went through 200. We still have about 10 to go this morning and after that we're going to go through the existing one, which a hundred-ish. Yeah, I still think that's too many. We need to go down.

Neal Ford:

Camilla is always sweating at this time during the face to face meeting because she's starting to get worried because the conversations are long. And so I want to go back to something you said earlier. You said you start, as you go through the wall, you start in the adopt ring and you always end in the hold ring. Is there a reason that you always start in the middle and go out toward hold or is that just something that's left over?

Camilla Crispim:

Yeah, and no particular reason I guess. Well when we are going through the new ones, adopt is always very easy to start with. So basically we only have moved there, and then trial and assess is what we see more blips, assess especially. Yeah, no particular reason. I don't know if there is a background history that I'm not aware.

Neal Ford:

I think it's because, I have a theory. So if you have, if you look at amount of conversation per blip per ring, which ring do you think dominates, given it usually has a very few number of blips in it, but has a large amount of conversation, is always the hold ring. And that I think is the reason we've always reserved that to last because it's like "Okay, these are going to drag on forever."

Camilla Crispim:

Makes sense.

Neal Ford:

If you start with that you'll never get anything.

Camilla Crispim:

Right.

Marco:

You'll be retired.

Camilla Crispim:

Yeah, that makes sense. That make sense.

Neal Ford:

I don't know if that was intentional at some point in the past or if it just sort of, you know, a kind of implicit observation that "Hey, if we want to get through this thing we should probably start here and work our way out."

Rebecca Parsons:

Then maybe people will be tired when they get to the hold ring.

Marco:

Yeah.

Neal Ford:

So how did you squeeze two weeks out of this process over the course of two years? Because that's pretty good. You knocked a big chunk out of an eight week process to get it down to six weeks.


Camilla Crispim:

Yeah. I think we started like slowly from like eight to seven. This time I was wishing to do seven to eight, but then we had some sort of like executive decision that came from Rebecca saying, "I think it's doable. I would really like to get it done in six weeks and let's see if we can make it happen." And I think it's actually doable. I think all the iterative sort of feedback and process that we are going through when we are writing the blips, getting feedback, so on, actually make that possible. So it's not a waterfall process for blips write up and feedback, but it's more-


Marco:

In small batches, right?


Camilla Crispim:

Small batches. Exactly.


Neal Ford:

Yeah, that was one of the things I think we started doing is once we get the blips reviewed we start sending with translation rather than getting them all done and then sending it like a big giant stage, which is much more iterative. And I think that probably helps the translators a lot.


Rebecca Parsons:

Well and even the review process is now iterative. It used to be we waited until all of the blips were written and then sent the entire group of blips to our internal community. And I actually think we also get better feedback because we send them in smaller batches and you're more likely to get people to spend a little time reading a half a dozen as opposed to the entire radar.


Neal Ford:

Yeah. The whole incremental change thing has legs, not just in software, I mean in lots of different things that have process.


Camilla Crispim:

Yeah, for life.


Neal Ford:

So you're wrapping up your two year tenure, which I'm sure is bittersweet because you don't have to come herd us twice a year anymore.


Camilla Crispim:

Which is very sad.


Neal Ford:

So if you have to give some advice to Nee, who's taking over your position, what would her number one piece of advice be? Besides learning how to ignore us?


Camilla Crispim:

I think she's already doing a pretty good job in feeling comfortable with the group. I think that's the big sort of thing, because sometimes you feel like, "Oh my God, the big thinkers of ThoughtWorks and the industry, I've read their books, like how come I'm here?" So there's this feeling. But I think she's already doing a good job on that.


Camilla Crispim:

The other thing is don't be shy to push people, because they need you to push them in order to get things done. And that's why you are there. They actually value that very much, and it's a very good thing, and it's another, it's a different perspective. It's no longer you getting it done yourself, but it's you being an enabler of getting it done and it's you succeed as people succeed. So that's this shift in the mindset, which is quite delighting.


Neal Ford:

Well, I think that's a fantastic way to wrap up. Thank you very much for joining us. She had to take time out of her busy day putting the radar together to do this. So we thank you very much for your chunk of time, and all the great work you've done on the radar.


Camilla Crispim:

Thank you.


Rebecca Parsons:

Thank you very much.


Neal Ford:

And we'll see you next time.

Check out the latest edition of the Technology Radar

More Episodes
Episode Name
Published