7 Why Build Revisited
Why Build… Revisited
Increased Reliance on IS Service Providers
In the Why Build… chapter we described how the supply of IS services to buy or rent from IS service providers is increasingly becoming a reasonable alternative for building systems or system components oneself. We mentioned savings in development and maintenance costs and the fact that such external services are typically more reliable and have been much more extensively tested than one’s own systems as good reasons to consider buying or renting rather than building.
We also noted that even those who build their own systems rarely build them from the ground up. Just like people who like building their own furniture typically buy their lumber and woodworking tools —of course, some woodworkers grow their own lumber and most woodworkers make at least some tools themselves— system developers typically do not build their own compilers or interpreters, operating systems, database software, web servers and web browsers, etc.
Still, when we compare the presence of ‘built’ rather than ‘bought or rented’ components in TE 2.0 with those of TE 1.0, we observe a significantly increased reliance on outside service providers. As information systems become more complex in that they provide more functions and as these functions often involve networked and Internet services, system builders increasingly ‘outsource’ these functions to external service providers. In this (very short) chapter) we again contrast TE 1.0 with TE 2.0, but this time from the perspective of build vs. rent.
TE Searches: From Tool To Service
The chapter on Searching TeachEngineering laid out the various approaches that were used in letting users search the TE collection of documents. Both TE 1.0 and 2.0 used their internally developed codes to search in their databases (tools) for faceted searches; e.g., grade-based searches or standard-based searches. TE 1.0 used a relational database, TE 2.0 a JSON database. However, whereas at first TE 1.0 used a locally installed tool (Lucene) for full text searches, both TE 1.0 and 2.0 ended up relying on external text search services: Google Site Search (now discontinued) in TE 1.0 and Microsoft’s Azure Search in TE 2.0. Whereas Google Site Search pricing was based on the number of searches conducted, Azure Search pricing is based on the amount of data stored and indexed.
Both TE 1.0 and 2.0 relied on the services of D2L’s Achievement Standards Network (ASN) for information about K-12 educational (science) standards. ASN’s cloud services let users access anything ASN knows about educational standards. Pricing is a fixed annual fee.
Both TE 1.0 and 2.0 rely on Google Analytics for analysis of their web traffic. Whereas TE 1.0 also maintained its own log of all significant website interactions, TE 2.0 no longer stores that information.
As discussed in a previous chapter, rather than using a traditional relational database such as MySQL as we did in TE 1.0, TE 2.0 uses the RavenDB document database. Also in TE 1.0 we chose to install, run, and maintain our own MySQL instance. Doing so, however, implies assuming responsibility for securing, upgrading, and supporting the database which is a significant burden for a small team. In TE 2.0, therefore, we are procuring RavenDB services from RavenHQ, a company affiliated with the developers of RavenDB, to provide us with a fully managed instance of RavenDB. RavenHQ takes care of database updates, security and the server infrastructure required to run RavenDB. Pricing is based on the desired performance level and amount of storage needed and includes the license fee for RavenDB itself.
Whereas TE 1.0 had essentially no social networking aspects, social networking has become an essential and quickly growing component of TE 2.0. The freely available AddThis widget allows users to share TE content via various social media channels such as Twitter, Pinterest, Email, and Facebook.
One very special category of TE users consists of the many authors who have contributed curriculum to TE. As described in the chapter on Document Accessioning, in both TE 1.0 and TE 2.0, author submissions are stored and managed in Open Journal Systems (OJS), an open source system for managing peer-reviewed journals. However, whereas under TE 1.0 the TE team ran its own installation of OJS, in the summer of 2017 TE 2.0 switched to using a cloud-based OJS provider. The main reason was that since OJS was and is used purely as a tool for managing curriculum submissions, it was considered cost effective to have a third party run and maintain it. The service follows an annual subscription model with pricing based on the amount of used storage.
Those who use TE as a source of teaching materials comprise, of course, a much larger group. Such users fulfill an important role in that they provide feedback on the functionality of the system, but also on the content of the information served by it. As such, it is in the mutual interest of both users and system maintainers to capture user feedback and to act on that feedback. Of course, some feedback is utter nonsense, disingenuous (refer to Appendix C: Fake Link Requests) or consists of nothing but a rant. Much more often, however, user feedback provides an angle on the system’s functionality or its content which was missed by its creators or maintainers and which, when taken into account, can lead to functional or content improvements.
TE 1.0 had a simple, internally developed Contact Us web page by means of which users could submit any sort of comment or question. These comments were internally processed by the TE team and were not shared on the TE site. Users could also submit reviews of individual lessons or activities through an internally developed Teacher Reviews web function. Unlike the Contact Us feedback, Teacher Reviews were shared on the site.
In TE 2.0 the Contact Us facility was retained but other user involvement functions were greatly expanded. One encompassed switching from the internally hosted Teacher Reviews facility to LinkEngineering, a cloud-hosted community platform where K-12 educators can share engineering experiences. LinkEngineering is a collaborative project of the US National Academy of Engineering and several other engineering-related organizations and is funded by Chevron. As part of an agreement with one of TeachEngineering’s funders, we decided to drop our own internal Teacher Reviews facility and replace it with the LinkEngineering one.
A second change was the deployment of the Disqus commenting service, hosted by , on the TE 2.0 curriculum pages. The Disqus service collects user feedback in the form of comments and allows other users to search, add and react to those comments with their own, and connects users and their comments with social network platforms. Disqus is free for non-profit organizations.
To further expand our engagement with users, TE 2.0 also uses the email marketing automation platform MailChimp for maintaining a mailing list of users who expressed an interest in receiving our newsletter. MailChimp allows us to create and send email newsletters and other email messages. MailChimp is free for up to 2,000 mailing list subscribers.
Comparing TE 1.0 with TE 2.0, the trend of increasing reliance on outside service providers is clear. Whereas previously we built our own, working but relatively primitive and often buggy services, we nowadays increasingly rely on far more comprehensive and more stable externally provided services. These services not only work better, but they give us more functionality and are typically less expensive than building and maintaining our own.