開源日報每天推薦一個 GitHub 優質開源項目和一篇精選英文科技或編程文章原文，堅持閱讀《開源日報》，保持每日學習的好習慣。

今日推薦開源項目：《別人家的 HTML 和 CSS(Ver4.0) purecss-pink》
今日推薦英文原文：《Will The Open Source Movement Backfire?》

今日推薦開源項目：《別人家的 HTML 和 CSS(Ver4.0) purecss-pink》傳送門：GitHub鏈接

推薦理由：是的你沒有看錯，那個作者他又出新作了。如果你是第一次看到這樣的標題的話推薦你去看看工場里這位作者之前的作品，拿 CSS 畫畫還能畫的這麼好的人可不多見。具體是什麼樣的就請各位去項目里看一看吧，別忘了如果有 Chrome 的話記得用那個打開。

今日推薦英文原文：《Will The Open Source Movement Backfire?》作者：Bill Franks

原文鏈接：https://medium.com/analytics-matters/will-the-open-source-movement-backfire-5dc9e661b2c3

推薦理由：開源行為會適得其反嗎？這個問題的角度非常新穎，就讓我們來看看作者是如何認為的。

Will The Open Source Movement Backfire?

While open source software has been around for decades, its adoption level and breadth of focus have exploded in recent years. Most companies now make regular use of open source software — and that usage is increasing. However, there is a potential downside to the open source explosion that may be heading our way. Is it possible that the open source movement will backfire on us? This blog will explore that possibility.

OPEN SOURCE STARTED FOCUSED

In the early days of open source software, there was heavy focus on a smaller number of large projects that were broadly adopted. Web browsers and web server software (Apache) were two early and popular open source focal points. Both of these areas are now dominated by open source. There is no turning back from open source in areas like these since the software is an integral part of so many consumers』 and companies』 processes.

This isn』t to say that a wide range of open source projects didn』t exist years ago. But, only a small set of open source projects made it into the hands of a broad swath of the population. This focus made it easier to get contributors to keep the software up to date and to add new features.

OPEN SOURCE TODAY IS WIDELY SPREAD

Open source is no longer focused only on broad functionality that the masses require. There are now also open source projects focused on targeted, niche areas. Within the analytics world, for example, we have R and Python that cover a broad range of analytic capabilities. But we also have niche projects that address only certain aspects of the analytics process. These include projects like D3 for visualization and Scikit-learn for machine learning libraries. On the technology and operating system side, there are a wide array of options available ranging from Hadoop, to Spark, to PostgreSQL.

The point is that there are now multiple open source projects focused on identical subject areas. As opposed to a single analytics-focused open source project, there are dozens or hundreds. Companies are also adopting and incorporating multiple of these projects into their internal processes. This works fine as long as the community keeps the software up to date. However, it can become a big problem if some of the projects end up abandoned or mostly ignored by the developer community that 「owns」 it.

HOW OPEN SOURCE MIGHT BACKFIRE

There are only so many top-notch programmers who know a given subject area well and who also have time they are willing to devote to contributing to open source projects. As more and more projects open up, the pool of available contributors is stretched further and further and they are forced to focus on a few projects they are most passionate about.

This is where the potential problem comes in. A few years down the road as many of the long-standing Apache HTTP server contributors retire, will younger programmers have interest in taking up an 「old, unexciting」 project like Apache HTTP server? They may instead opt to contribute to sexier options like TensorFlow in the AI space. This means that even mature, widely deployed open source software could struggle to maintain support over time.

Worse, will the new projects mature effectively? With developers hopping from project to project, open source could become like the nightclub scene. The hot club this month with all the crowds could be dead in six months as a new club takes over the market. The new club gets all the attention for a while until yet another club takes over.

For example, Hadoop was among the hottest, sexiest open source projects out there for a few years. There was no problem getting world class contributors to help Hadoop grow and expand very quickly. Today, however, Hadoop is becoming widely regarded as yesterday』s news. Much of the contributor community may start to move on to other, sexier projects. This does not bode well for Hadoop』s ability to maintain enterprise readiness in the long term. Keep in mind that Hadoop is simply an illustrative example of the broader issue. I』m not trying to pick on Hadoop.

HEDGING YOUR BETS

Not long ago, many organizations were hesitant to implement open source tools at all. Further, when open source was implemented, it was at a strategic, enterprise level. Today, most organizations have embraced open source and are allowing groups such as analytics organizations to implement open source tools specific to their domain. If that hot new open source analytical tool doesn』t take off as expected, an organization will be stuck with orphaned software that has no support structures in place.

Much of the open source software being implemented today is far less mature and far more niche than implementations of the past. Are we heading for a future filled with a lot of half-completed, unsupported, low quality open source projects in place? Possibly so.

The action to take is to be careful and deliberate in your pursuit of open source toolsets. Don』t just install every up and coming project and start building critical processes with it. Focus on implementing mature projects with a wide contributor base and make sure the project looks like it will be here for the long haul. Then, track and monitor the state of any implemented open source tool carefully so that there is time to react if the project looks to be heading for trouble. Finally, build your processes in such a way (like a Lego kit!) that it is easy to swap out components without disrupting the entire chain.

The last thing anyone wants is to be caught with a mission critical piece of software that suddenly needs to be replaced. The software may be free but fixing the problem will be far from it.