‘Readers respond’ is a forum for individuals to contribute their own reactions, perspectives, or experiences in relation to a specific article published in a recent issue. These views are not necessarily those of the Editor, the Editorial Panel, or the Publisher. This article responds to Gan, Z. and C. Leung. 2020. ‘Illustrating formative assessment in task-based language teaching’. ELT Journal, 74/1: 10–19. E-portfolios are broadly defined as digital dossiers, where students create, connect, showcase, and reflect upon multimodal artefacts to enhance learning (Yancey 2009). In the past two decades, they have been widely applied in large-scale standardised testing and classroom-based assessments, particularly of L2 writing. E-portfolios have multifaceted advantages. They can improve students’ writing performances, motivation, and self-regulated learning (Aygün and Aydin 2016). They can promulgate process writing, self- and peer assessment, and self-reflection. They can also be used to evaluate higher-order thinking skills, namely creativity, problem-solving skills, or metacognition, which would not be easily measured by one-shot essay testing. Despite these, e-portfolios have limitations. They require students and teachers to possess high to moderate levels of computer literacy. Accessibility to infrastructures like Wi-Fi connection, electronic gadgets and portfolio software would be causes for concern, especially for those economically less advantaged students. The issue of whether assessing writing performances or assessing technological skills remains contentious, given that teachers are unsurprisingly distracted by flashy e-portfolio designs (Barrett 2007). In reality, teachers tend to use e-portfolios to fulfill multiple assessment purposes, such as assessment of learning (AOL) and assessment for learning (AFL) concurrently. The former is to evaluate learning, whereas the latter is to support learning. Using one tool to achieve numerous assessment goals seems to be promising, but requires specific knowledge, skills, and commitments to make it happen. Scholars have long advocated combining AOL and AFL to promote positive learning (Taras 2005), but evidence has shown that teachers were unable to manage both assessment purposes simultaneously owing to a host of individual (teacher misconceptions), institutional (restrictive school cultures), and contextual (exam-driven ideologies) barriers (Gan and Leung 2020). Further, there is a complex but very important relationship between teachers’ view/s of teaching and learning, and their assessment approach/es. The point here is that e-portfolios can be used in different ways by teachers, much depending on their pedagogic beliefs and values (James 2006). Against this backdrop, I argue that it is not necessary to integrate AOL and AFL both at once. Instead, I propose that AFL occurs during the processes of e-portfolio construction, whereas AOL takes place at the end of term. This idea is not meant to be rigid, but assumes that e-portfolio assessment data serve different purposes at different times. The rationale for separating AOL and AFL is threefold. First, teachers may find it manageable to conduct AFL before AOL rather than AFL+AOL/AOL+AFL in a blended mode. Ultimately, AFL should benefit AOL, because teachers provide students with gradual, scaffolded instructional guidance before their e-Portfolios are graded. Second, assessing the e-portfolio compilation processes is a tall order, because it is exacting to evaluate the process-oriented and developmental nature of L2 writing dependably. Third, the segregation of AOL and AFL causes least disruption to teachers’ everyday pedagogical agendas. Here, I highlight the five main issues teachers may encounter when they assess the formative aspects of e-portfolios summatively throughout a school term/year. First, it is labour-intensive to monitor students’ portfolio compilation, let alone providing students with ongoing feedback on their interim drafts and other multimodal artefacts like vlogs. Second, teachers may find it problematic to assess diverse genres with one holistic rubric no matter how rigorously it was constructed and validated, as students’ e-portfolios involve complex content knowledge and compilation skills. Third, ethical issues remain a concern when teachers require students to upload their e-portfolios on public domains. The infringement of privacy and confidentiality could be alarming, especially when students confess their incompetency in learning and expressing their emotional responses to teachers’ marking. Fourth, teachers may find it difficult to distinguish whether students’ reflective pieces are authentic, because students could mask their self-criticisms with implicit self-enhancement strategies (McGarr and O’Gallchóir 2020). Fifth, because e-portfolios emphasize collaboration, identifying students’ authorship would become taxing when teachers assess individual writing performances for coursework marks. To promote AFL during e-portfolio compilation, teachers can ask students to do basic research on a topic, and create their e-portfolios by a free Web 2.0 application, Wix (https://wix.com) or an e-portfolio tool, Seesaw (https://web.seesaw.me/). Teachers encourage students to be creative to develop e-portfolio contents. Then, students are required to connect their multimodal artefacts, such as audios, videos, texts, graphics, and hyperlinks. They collate, curate, and showcase these online texts both chronologically and thematically to take stock of their learning. Lastly, assisting students to reflect upon their artefacts is most crucial. Teachers could provide students with guided questions or related vocabulary items when students perform self-reflection. As mentioned above, these formative aspects of e-portfolios are creative, process-oriented, and reflective, which are somewhat demanding to be assessed summatively. Rather, giving students formative feedback on their artefacts to close their learning gaps makes more sense and dovetails with the theoretical principles of e-portfolios—AFL—using assessment to support learning rather than to evaluate learning (Barrett op.cit.; Black 2015). In primary and secondary schools, teachers must grade students’ e-portfolios for accountability purposes. They score students’ final products—a completed e-portfolio—with a rubric, including process, product, and reflective elements. The weighting of the process grade could be set as 30 per cent to reward students’ involvement in conferencing, efforts in creating/revising artefacts, and commitments to portfolio compilation, such as organising hyperlinks. The weighting of the product grade could be set as 50 per cent according to the various aspects of e-portfolio writings, namely creativity, logical thinking, organisation, rhetoric, accuracy, clarity of multimodal texts, etc. Teachers are advised to use one analytical rubric to assess four or five artefacts to enhance the scoring validity of e-portfolios. The weighting of reflective pieces is most problematic, as this unique genre should not be graded. Teachers may allocate 20 per cent to grade the evidence in support of the statements presented in the reflective pieces. This evidence could be in the form of scores, completed works, peer/teacher feedback or any artefacts that are indicative of students’ performances. Although researchers recommend synergising AOL and AFL in e-portfolios, teachers may find it impractical to do so owing to their misunderstandings of AFL and lack of skills in integrating AOL and AFL effectively. To echo my initial argument, teachers need not reinvent the wheel by mastering eclectic instructional strategies to fulfil all assessment purposes at one go. Alternatively, teachers’ judgments in AOL should be prioritised so that valid and dependable assessment evidence can be generated to inform teaching and student learning (Assessment Reform Group 2006). AFL and AOL could be separated to perform their respective formative and summative functions to the full. After all, e-portfolios are originally designed for learning. In sum, albeit not mutually inclusive, AOL and AFL still have divergent roles to play in tracking and documenting students’ learning in e-portfolios.