In­ter­view

Linux User & Developer - - Contents - nate Graham Nate, a for­mer Ap­ple engi­neer, is a KDE de­vel­oper and prod­uct man­ager for Kubuntu.

KDE de­vel­oper, Nate Graham on the project’s bug-fix­ing blitz

KDE is a vast project that cel­e­brates 22 years in Oc­to­ber. We spoke to nate Graham about KDE’s on­go­ing bug-fix­ing charm of­fen­sive

KDe de­vel­oper nate Graham spear­heads KDe’s usability and pro­duc­tiv­ity ini­tia­tive, which is one of the three ma­jor goals of the or­gan­i­sa­tion that were agreed in 2017.

“The thrust of this goal is we want to make KDE Plasma and our KDE apps eas­ier to use, and more ap­proach­able for typ­i­cal users,” Graham told us, chat­ting over VoIP from his sofa at home – and the re­sponse, par­tic­u­larly for re­cent Plasma point re­leases, has been pos­i­tive. Go­ing fur­ther, KDE wants to make com­plex work­flows a lot simpler and eas­ier for any­one to use: “We want our soft­ware to be a lean, mean pro­duc­tiv­ity ma­chine,” says Graham. Read our, (lightly edited) in­ter­view be­low:

Let’s start with what you do on the KDe project…

I write doc­u­men­ta­tion to help peo­ple get in­volved in the project. I do bug triag­ing, so that we can have bug-track­ing sys­tems that are ac­cu­rate. I men­tor for new con­trib­u­tors to help them get in­volved in this; I sub­mit lots of patches my­self. I re­view patches that other peo­ple have sub­mit­ted. I do user in­ter­face de­sign, and usability and vis­ual stuff. I’m in­volved in this process quite a lot, and my work takes me into con­tact with lots of other mem­bers of the KDE com­mu­nity, which is one of the most re­ward­ing parts of it, be­cause I get to learn from them, and I hope that they get to learn from me, too.

I won­dered how you ended up work­ing on KDe from work­ing at Ap­ple?

It’s a funny jour­ney. Work­ing at Ap­ple was al­ways one of my child­hood dreams. I al­ways loved Ap­ple’s soft­ware and hard­ware, be­cause I felt that they re­ally fo­cused re­ally ag­gres­sively on both the over­all user ex­pe­ri­ence and also the lit­tle pol­ished de­tails that were of­ten over­looked in other soft­ware. They re­ally said, “It’s not okay that we get 95 per cent of the way there. We need to get 100 per cent of the way there.” I worked at Ap­ple for eight years, but even­tu­ally, it was time to move on. Once that was done, I de­cided that I wanted to be in­volved in a part of the Linux com­mu­nity.

So I looked around for a lit­tle while, and I tried to find a Linux project that I felt matched the most what I saw Ap­ple was do­ing. And I found a num­ber of dif­fer­ent ones. But ul­ti­mately, the rea­son why I set­tled on KDE was not so much I felt they were cur­rently the most Ap­ple-like, but be­cause I felt like KDE’s com­mu­nity was the one that was most wel­com­ing to­wards be­com­ing this way – and the most wel­com­ing to­wards me, as well.

You put in your usability and pro­duc­tiv­ity pro­posal last year. You’ve cov­ered it a lit­tle bit, but could you tell us a lit­tle more about the ini­tia­tive?

When I ini­tially pro­posed this and it was ac­cepted, I set a time­line for two years. I said to my­self,

‘I think we need to do ba­si­cally 10 years’ worth of im­prove­ment ev­ery year.’ So 20 years’ worth of prod­uct usability im­prove­ments in two years, maybe 30 in three years. But I think we can ab­so­lutely get there and as the ex­cite­ment for this ini­tia­tive has built up, I think we see an in­creas­ing num­ber of peo­ple want­ing to be­come in­volved in it, which is re­ally re­ward­ing – be­cause lots of new con­trib­u­tors means lots of new help, and the pace of the change hap­pens faster. Which makes ev­ery­body else ex­cited, and it be­comes a vir­tu­ous cir­cle.

Has that led to changes in qual­ity test­ing? Has it changed how you test the apps and frame­work? The way we do our test­ing is some­where I feel like we still have some im­prove­ments to make. This is sort of an en­demic prob­lem ev­ery­where, be­cause test­ing is one of the least fun as­pects of soft­ware de­vel­op­ment, of course. And all of our soft­ware has very ag­gres­sive re­lease sched­ules – we re­lease three new ver­sions of Plasma ev­ery year. We re­lease three new ver­sions of all our apps ev­ery year. And

we re­lease a new ver­sion of the frame­works that sup­port all of them once a month – so 12 times a year. Of­ten, this doesn’t leave enough test­ing time.

For frame­works, there’s al­ways the pos­si­bil­ity that if some­thing breaks or if there’s a re­gres­sion, we can just ship a fix the next month. But for all the other ones, it means that we have very ab­bre­vi­ated re­lease sched­ules, which means that the de­mand for testers and test­ing is al­ways quite ex­treme. One of the things we’re re­ally look­ing to do is to help peo­ple test the beta ver­sions of our soft­ware.

We al­ready have quite a num­ber of testers who do this. Those peo­ple are amaz­ing. I want to give a shout out to them be­cause they are so im­por­tant and awesome. But we can al­ways use more help. Test­ing the beta ver­sions of KDE soft­ware is prob­a­bly one of the big­gest ways that a non­soft­ware de­vel­oper con­trib­u­tor can help to keep the qual­ity lev­els up.

You’ve had quite a fo­cus on Dis­cover re­cently? Yes, I feel Dis­cover has made huge strides in the last few re­leases. When I started with Plasma, the feed­back that I al­ways saw with Dis­cover, at least on­line from users, was quite neg­a­tive – and of­ten quite hos­tile, which was a lit­tle bit de­mor­al­is­ing.

That was one of the rea­sons why I felt we re­ally needed to fo­cus on this project. It wasn’t that it was a bad project: there were a few lit­tle nig­gling is­sues of the sort that tend to an­noy users. Even if the foun­da­tion is re­ally sta­ble, if the in­ter­face isn’t pol­ished to the ex­tent that peo­ple want or it’s not to­tally re­li­able, peo­ple of­ten find they’re frus­trated.

So I worked with the lead de­vel­oper, Aleix [Pol], who is an amaz­ing guy, and has been in­volved with KDE for a very long time. He, him­self, had no­ticed most of these is­sues and was al­ready mov­ing to­wards re­solv­ing them. So I helped out, mostly on the user in­ter­face side, to make Dis­cover look and feel more like a Desk­top app when it’s run­ning on the Desk­top, and to solve some user in­ter­face in­con­sis­ten­cies, and to pro­pose some bet­ter be­hav­iours, and to re­design some of the pages to look more vis­ually at­trac­tive.

I think we saw a pretty good re­sponse for the

5.12 re­lease. The 5.13 re­lease, we got a re­ally good re­sponse. And the up­com­ing 5.14 re­lease, I think peo­ple are go­ing to like it even more, be­cause we’ve made even more of the kinds of changes that they want. And now Dis­cover is sta­ble. It has lots of fea­tures. It’s very vis­ually at­trac­tive. So per­son­ally, I think Dis­cover is an ex­am­ple of a cur­rent suc­cess story, where we fo­cused a lot of our re­sources ag­gres­sively on one app that needed some at­ten­tion. Now that we’re there, we can re­di­rect those re­sources else­where, and move onto the next thing.

It’s a very large project, so how do you deal with the in­ter­op­er­abil­ity and com­pat­i­bil­ity is­sues with KDe in terms of de­pen­den­cies and con­sis­tency? KDE Plasma and KDE apps are run on so many dif­fer­ent dis­tros, and some peo­ple even run some of them on Win­dows and Mac op­er­at­ing sys­tems. We of­ten do have lots and lots of bug re­ports from users. To a cer­tain ex­tent, this hap­pens in lots of com­puter soft­ware, be­cause de­vel­op­ers tend to use cer­tain

I want to give a shout out to our testers be­cause they are so im­por­tant and awesome

types of sys­tems, and users tend to use a vastly more di­verse set of hard­ware and soft­ware than we do. This is why hav­ing beta testers is so im­por­tant.

I want to re­ally thank the users of rolling re­lease dis­tros at this point, be­cause to a cer­tain ex­tent, in the whole Linux ecosys­tem, they pro­vide a re­lease QA for ev­ery­body else. This is, I think, es­pe­cially true for KDE soft­ware, be­cause of our ag­gres­sive re­lease sched­ules. Ev­ery sin­gle month, we re­lease a new ver­sion of KDE frame­works. So that gets into the hands of Arch users prac­ti­cally im­me­di­ately.

So if we get a whole bunch of bugs from Arch users that say, ‘Hey, this broke’, we can fix it in the next re­lease of KDE frame­works. And that’s go­ing to be com­pletely in­vis­i­ble to the peo­ple who use Kubuntu, who use De­bian, who use openSUSE. They ba­si­cally ben­e­fit­ted from the Arch peo­ple’s QA. So I’d just like to say thank you, Arch and Man­jaro and Tum­ble­weed users. You’re all do­ing the com­mu­nity, and the wider Linux ecosys­tem, a re­ally in­cred­i­ble ser­vice.

what do you feel are the bar­ri­ers for KDe?

I gave a pre­sen­ta­tion at Akademy [KDE’s com­mu­nity gath­er­ing], and I tar­geted a num­ber of user per­sonas whom I think can ben­e­fit from KDE Plasma. For the most part, these are tech­ni­cal pro­fes­sion­als, busi­ness pro­fes­sion­als, stu­dents and teach­ers, peo­ple along those lines.

I think what we need to do is look at the rea­sons why they don’t cur­rently use our hard­ware and soft­ware. A ma­jor one is a lack of pre-in­stal­la­tion on peo­ple’s hard­ware. I think that re­ally is the big­gest one. Be­cause most peo­ple don’t buy soft­ware; they buy hard­ware. To most peo­ple, the phys­i­cal thing they can hold in their hands is the prod­uct. So we need to get KDE Plasma and KDE apps on more hard­ware. But in or­der to do that, what we need to do is first fo­cus on our soft­ware to the point where hard­ware ven­dors can look at it and say, ‘Hey, wow, this thing is so good. I re­ally want to put this on my hard­ware.’ And then it’ll be at­trac­tive to users. So I work back­wards from there. I say, ‘We’ve got to put it on hard­ware. To do that, we have to fo­cus on soft­ware. But which ways do we fo­cus on soft­ware?’

We need to make sure that it’s meet­ing busi­ness and en­thu­si­ast power-user use cases, and that it’s meet­ing stu­dent and teacher use cases. Those are the pri­mary ones.

From that very high level per­spec­tive, that leads to some tech­ni­cal things. The peo­ple who I’m talk­ing about very fre­quently use re­mote files on Samba shares, for ex­am­ple. This is some­thing that tech­ni­cally we’re a lit­tle bit be­hind on. Our tech for ac­cess­ing files and fold­ers on Samba shares is not a re­li­able as it needs to be. So that’s some­thing I would re­ally like to fo­cus on. I feel like if we man­aged to do that, we re­duce the fric­tion a lit­tle bit for the peo­ple in those use cases. Then we make the soft­ware more at­trac­tive, and the hard­ware ven­dors are go­ing to want it, and then peo­ple are go­ing to be able to buy stuff with it on. And then we take over the world…

As prod­uct man­ager for Kubuntu, how’s 18.10 com­ing along?

Pretty well. We’re go­ing to have Qt 5.11. We’re go­ing to have Plasma 5.13. We’ve got­ten a very good re­cep­tion from Kubuntu 18.04, users are pretty happy with it. We’ve made a cou­ple of very bold user in­ter­face and de­faults de­ci­sions, too, that I was hop­ing peo­ple would re­spond well to – and so far, they have. This in­cluded one that I wasn’t fully on board with orig­i­nally, which was us­ing a dark theme

for Plasma and a light theme for apps. That was some­thing I fought against in­ter­nally, but now it’s been re­leased, peo­ple re­ally seem to like it. So I’m go­ing to say: I was wrong, you were right – that was a great de­ci­sion. That’s one of the things that I think makes our com­mu­nity re­ally strong. We have a lot of peo­ple who are in­ter­ested in push­ing the en­ve­lope.

what do you feel were the key de­vel­op­ments from Akademy this year?

That’s a re­ally great ques­tion. I think there were a cou­ple of things that were re­ally im­por­tant. First of all, the Kate and KTex­tEdi­tor de­vel­op­ers did prob­a­bly a year’s worth of work in one week, which was as­ton­ish­ingly amaz­ing. It was an in­cred­i­ble pace of im­prove­ment. Other than that, there was lots of work on Plasma, too. There were lots of re­ally great con­ver­sa­tions that we all had re­gard­ing top­ics that are some­times hard to dis­cuss over the in­ter­net but are a lit­tle bit eas­ier to have face to face. There isn’t re­ally any­thing spe­cific that I’d like to talk about just yet, but we got rid of a lot of bot­tle­necks.

So for de­vel­op­ers in­ter­ested in get­ting in­volved, what kind of sup­port does the KDe project of­fer? We have a wiki page that’s linked from the main page on KDE.org. Just click ‘Get in­volved’ at the very top, and it takes you to a whole list of things that you can do. We’ve got de­tailed doc­u­men­ta­tion and in­struc­tions for how to file good bugs, how to triage bugs, how to get in­volved in de­vel­op­ment, how to get in­volved in vis­ual de­sign and trans­la­tion. We’re al­ways work­ing to im­prove that doc­u­men­ta­tion, but we think it’s at a pretty de­cent stage right now.

Once you de­cide that you want to get started, there are a lot of great re­sources for com­mu­ni­cat­ing

with de­vel­op­ers. We’ve got IRC chan­nels, we’ve got Tele­gram rooms, we’ve got mail­ing lists and then of course, our sys­tem for code re­view and test man­age­ment – Phabri­ca­tor [https://phabri­ca­tor. kde.org] is some­where where we’re all very ac­tive.

On pri­vacy, what de­vel­op­ments are hap­pen­ing? The pri­vacy side is not as ac­tive as the other two [goals], but there are def­i­nitely some things go­ing on be­hind the scenes. There’s been a lot of prepa­ra­tion for work that hasn’t been made pub­lic yet.

Is that re­lated to Purism’s Li­brem 5 phone?

I think it’s re­lated, in the sense that all of these things are kind of re­lated, be­cause we’re do­ing lots of work with them. I think when it comes to the Li­brem 5 phone, we want to make our plat­form more at­trac­tive for that tar­get. I think one of the ways we’re do­ing that is with the GNU Ring [a uni­ver­sal com­mu­ni­ca­tion plat­form fo­cused on pro­tect­ing pri­vacy and free­doms]. This is the KDE client for the GNU Ring soft­ware. We’ve just re­leased a new ver­sion re­cently that uses our Kirigami frame­work in or­der to im­prove the user in­ter­face. So we’ve just got to keep go­ing with that, and we re­ally hope that we can strengthen our part­ner­ship with [Purism], and even­tu­ally con­vince them to use Plasma Mo­bile.

we were un­der the im­pres­sion that Purism wanted to sup­port it?

My im­pres­sion is that [Purism] are, right now, go­ing with a cus­tom-made GNOME so­lu­tion. But they’re com­mit­ting to sup­port­ing a com­mu­nity-main­tained KDE Plasma ver­sion. So by de­fault, I don’t be­lieve they’re cur­rently plan­ning to ship with Plasma Mo­bile. That’s some­thing we hope to change.

We’ve got­ten a very good re­cep­tion from Kubuntu 18.04– users are pretty happy with it

Above Kubuntu’s 18.04 new dark theme for Plasma, with a light theme for KDE apps, turned out to very pop­u­lar

Above KDE has re­cently re­leased a Qt-based client for GNU Ring (https://ring.cx), the free and uni­ver­sal com­mu­ni­ca­tion plat­form. Ring uses dis­trib­uted hash ta­ble tech­nol­ogy to cre­ate its own net­work. Since July 2018, GNU Ring now sup­ports fea­turessuch as au­dio and video call record­ing and push no­ti­fi­ca­tions

Newspapers in English

Newspapers from UK

© PressReader. All rights reserved.