4 ms·

It's interesting to see the popularity of testing frameworks, continuous integration, and static code analysis in the Java community. What do people think abou

It's interesting to see the popularity of testing frameworks, continuous integration, and static code analysis in the Java community. What do people think about Gradle as a build tool (versus Maven)?

jebblue 12y ago
Personally I don't like the look of the code for Gradle. make, Ant and Maven I get, prefer Ant.
lmm 12y ago
Maven is wonderful because it means I can pick up any Java project and there will be no surprises in how it builds; it simply won't allow your project to be a special snowflake with its own weird rules. I wish the community would cherish this and stop writing tools that make it possible to set up more complicated builds.
- vorg 12y ago
  From the article... > In the past, Maven (64%) and Ant + Ivy (16.5%) have been more or less neck and neck. Indeed, we believe that many developers often use both in different ways on complex projects – in 2012, 67% of respondents used Maven and 48% used Ant. [...] Ant with or without Ivy is only one-third as commonplace as it was reported in 2012. The rapid increase of Gradle (11%) is also interesting to witness. Since 2012, Gradle use has more than doubled The 2014 question was what build tool is used most often, whereas the 2012 figures are for what build tool was used, with some overlap. So Maven use has stayed high for those using it, whereas Ant users have either given up on it, use Maven more often, or have switched to Gradle (which is still Ant). I wouldn't trust the download numbers given by the Gradleware CEO, though. The Groovy ecosystem has a long history of fabricating numbers, e.g. look at https://bintray.com/groovy/maven/groovy/view/statistics https://bintray.com/groovy/maven/groovy/view/statistics and see 865,000 downloads of Groovy in the past month, then click on country and see 822,000 of them are from China, and only 17,000 from the US.
MrBuddyCasino 12y ago
I've used Ant, Gradle and Maven - they all suck, but in different ways. Maven is too restrictive and its XML too verbose, but Gradle can turn to spaghetti code, its DSL is often hard to debug and sometimes under-documented, and could be faster imho. Maybe one day Gradle will suck less than Maven, but currently I think they're about the same.
- snambi 12y ago
  "gradle" is nothing but "ant", though not in XML. Gradle is appealing to the crowd that likes ant. Also gradle is super slow compared with ant or maven.
  - MrBuddyCasino 12y ago
    Actually no, Gradle is Ant + dependency management + reusable code (plugins) with a convention-over-configuration philosophy, using Groovy instead of XML. So no, Gradle is not just Ant. But god is it slow, I agree to that.
fauigerzigerk 12y ago
We don't know anything about the popularity of these tools in the Java community. This is not a random sample of Java developers. It's a self selected group of people who took part in a survey conducted by a Java tools vendor.
- karianna 12y ago
  The process was reasonably random - the survey was distributed to the global Java User Groups & I'd say over 2000 responses is something worth paying attention to statistically (yeah, I know - out of x Million Java developers it's still a tiny sample set).

It's interesting to see the popularity of testing frameworks, continuous integration, and static code analysis in the Java community. What do people think about Gradle as a build tool (versus Maven)?

jebblue 12y ago

Personally I don't like the look of the code for Gradle. make, Ant and Maven I get, prefer Ant.

lmm 12y ago

Maven is wonderful because it means I can pick up any Java project and there will be no surprises in how it builds; it simply won't allow your project to be a special snowflake with its own weird rules. I wish the community would cherish this and stop writing tools that make it possible to set up more complicated builds.

vorg 12y ago
From the article... > In the past, Maven (64%) and Ant + Ivy (16.5%) have been more or less neck and neck. Indeed, we believe that many developers often use both in different ways on complex projects – in 2012, 67% of respondents used Maven and 48% used Ant. [...] Ant with or without Ivy is only one-third as commonplace as it was reported in 2012. The rapid increase of Gradle (11%) is also interesting to witness. Since 2012, Gradle use has more than doubled The 2014 question was what build tool is used most often, whereas the 2012 figures are for what build tool was used, with some overlap. So Maven use has stayed high for those using it, whereas Ant users have either given up on it, use Maven more often, or have switched to Gradle (which is still Ant). I wouldn't trust the download numbers given by the Gradleware CEO, though. The Groovy ecosystem has a long history of fabricating numbers, e.g. look at https://bintray.com/groovy/maven/groovy/view/statistics https://bintray.com/groovy/maven/groovy/view/statistics and see 865,000 downloads of Groovy in the past month, then click on country and see 822,000 of them are from China, and only 17,000 from the US.

MrBuddyCasino 12y ago

I've used Ant, Gradle and Maven - they all suck, but in different ways. Maven is too restrictive and its XML too verbose, but Gradle can turn to spaghetti code, its DSL is often hard to debug and sometimes under-documented, and could be faster imho. Maybe one day Gradle will suck less than Maven, but currently I think they're about the same.

snambi 12y ago
"gradle" is nothing but "ant", though not in XML. Gradle is appealing to the crowd that likes ant. Also gradle is super slow compared with ant or maven.
- MrBuddyCasino 12y ago
  Actually no, Gradle is Ant + dependency management + reusable code (plugins) with a convention-over-configuration philosophy, using Groovy instead of XML. So no, Gradle is not just Ant. But god is it slow, I agree to that.

fauigerzigerk 12y ago

We don't know anything about the popularity of these tools in the Java community. This is not a random sample of Java developers. It's a self selected group of people who took part in a survey conducted by a Java tools vendor.

karianna 12y ago
The process was reasonably random - the survey was distributed to the global Java User Groups & I'd say over 2000 responses is something worth paying attention to statistically (yeah, I know - out of x Million Java developers it's still a tiny sample set).